AU2002348748A1 - Supply chain optimization - Google Patents

Description

WO 03/054756 PCT/IB02/05580 1 SUPPLY CHAIN OPTIMIZATION CROSS REFERENCE TO RELATED APPLICATION This application is based on, and claims the benefit of, co-pending U.S. Application 5 Serial No. 10/024,525, filed on 21 December 2001, entitled "Supply Chain Optimization," and incorporated herein by reference. BACKGROUND OF THE INVENTION FIELD OF THE INVENTION The present invention relates to supply chain management. More particularly, the present 10 invention relates to systems and methods for analyzing an inventory supply chain to optimize its efficiency and effectiveness. DESCRIPTION OF RELATED ART The channels that goods or resources travel through from production through first sale are known as a supply chain. These channels may extend from a manufacturing point to a retail 15 sales location or from a point where a resource, such as raw ore, is harvested to a manufacturing location, where a product is made from that resource. In the manufacturing process or sales process, over-supply or under-supply of goods or resources is undesirable. An efficient supply chain maintains the optimum amount of goods and resources throughout the supply chain to avoid both overstocking and under-stocking. Under-stocking may result in lost sales or slowed 20 down production. Similarly, overstock results in costs to store the excess stock, and may require sales at lower prices to move the excess inventory. Conventionally, the management of a supply chain is a difficult and complicated task. The complexity of the supply chain itself, as well as the large number of factors that may affect the efficient replenishment of product flow in the supply chain, makes optimization extremely 25 difficult. Therefore, there is a need in the art for tools that facilitate the efficient replenishment of product flow in a supply chain. CONFIRMATION COPY WO 03/054756 PCT/IB02/05580 2 SUMMARY OF THE INVENTION The present invention overcomes the problems and limitations of the prior art by creating a more efficient supply chain that maintains an optimum amount of goods and resources at various stages throughout the supply chain. This avoids both overstocking and under-stocking. 5 By avoiding under-stocking, lost sales and slowed production are minimized. By avoiding overstocking, stores reduce costs due to excess stock, and are not forced to sell goods at lower prices in order to sell excess inventory. The invention also facilitates the efficient management and manipulation of the many factors that may impact a supply chain, and facilitates the efficient replenishment of product 10 flow within the supply chain. Embodiments of the present invention optimize a supply chain so that a user may quantify factors that impact customer/consumer service and the available trade-offs. Aspects of the supply chain that have an appreciable impact on operational efficiency and effectiveness may be simulated and the results of supply chain simulations may be used to implement changes 15 in the actual supply chain and thereby optimize the efficiency and effectiveness of the supply chain. Embodiments of the present invention also achieve an efficient and streamlined supply chain by implementing strategies to link suppliers, retailers and consumers. In one aspect of the invention, there is a method of optimizing a supply chain that includes the steps of selecting 20 products representing segments of the supply chain, collecting supply chain data relating to the selected products, calculating demand for the selected products, developing a simulation of the supply chain to produce a baseline for the supply chain, and altering factors effecting the simulated supply chain to produce an optimized supply chain. In other aspects of the invention, the method may be embodied in a data processing system or as computer readable instructions 25 stored on a computer readable medium. Various embodiments of the invention may calculate the demand from sales and inventory data. The same or other various embodiments may provide graphical and numerical output in developing a simulation of the supply chain. Providing graphical and numeric output WO 03/054756 PCT/IB02/05580 3 may delineate actual sales, transportation movements, product inventory levels, and missed sales opportunities. The supply chain may be optimized in part by grouping similar stores and similar distribution centers in the supply chain. The altering of factors within the supply chain to 5 produce an optimized supply chain may also be used to develop future product flow paths and develop product development strategies. BRIEF DESCRIPTION OF THE DRAWINGS These and other attributes of the present invention will be described with respect to the following drawings in which: 10 FIG. 1 is a block diagram illustrating a computer system upon which aspects of the present invention may be embodied; FIG. 2 is a block diagram illustrating method of supply chain optimization according to an embodiment of the present invention; FIG. 3 is a block diagram illustrating the data formatting step shown in Fig. 2, according 15 to an embodiment of the present invention; FIG. 4 is a chart showing low and high profit scenarios; FIG. 5 is a main menu screen for the system and method of optimizing a supply chain according to an embodiment of the present invention; FIG. 6 is a screen for entering the names of the groups of stores in the supply chain 20 according to the present invention; FIG. 7 is a setup store group screen according to an embodiment of the present invention; FIG. 8 is a define demand screen according to an embodiment of the present invention; WO 03/054756 PCT/IB02/05580 4 FIG. 9 is a profiles screen containing a seasonal profile week used in the demand generation according to an embodiment of the present invention; FIG. 10 is a replenishment characteristics screen according to an embodiment of the present invention; 5 FIG. 11 is a replenishment characteristics menu according to an embodiment of the present invention; FIG. 12 is a parameters screen according to an embodiment of the present invention; FIG. 13 is an allocations screen according to an embodiment of the present invention; FIGS. 14a and 14b are replenishment methods screen according to an embodiment of the 10 present invention; FIG. 15 is a graph of the demand plot according to an embodiment of the present invention; FIG. 16 is a forecast methods screen according to an embodiment of the present invention; 15 FIG. 17 is a forecast parameters screen according to an embodiment of the present invention; FIG. 18 is a winters seasonal profiles screen according to an embodiment of the present invention; FIG. 19 is a replenishment days screen according to an embodiment of the present 20 invention; FIG. 20 is a forecasting screen according to an embodiment of the present invention; FIG. 21 is a graph showing the supply plot according to an embodiment of the present invention; WO 03/054756 PCT/IB02/05580 5 FIG. 22 is a results screen according to an embodiment of the present invention; FIG. 23 is a supplier availability screen according to an embodiment of the present invention; FIG. 24 is a winters screen according to an embodiment of the present invention; 5 FIG. 25 is a supply characteristics screen according to an embodiment of the present invention; and FIG. 26 is a block diagram of a computer readable medium according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION 10 The present invention may be used to create a more efficient supply chain that maintains an optimum amount of goods and resources throughout the supply chain to avoid overstocking and under-stocking of goods. By avoiding under-stocking, lost sales and slowed production are minimized. By avoiding overstocking, stores reduce costs due to excess stock, and are not forced to sell goods at lower prices in order to move the excess inventory. The invention may 15 also be used to efficiently manage and manipulate the many factors that impact a supply chain, and to facilitate the efficient replenishment of product flow within the supply chain. In order to provide solutions that create efficient supply chains, the present invention is preferably implemented in conjunction with one or more computers and one or more networks. For instance, aspects of the present invention may be embodied on a computer system, such as 20 the system 100 shown in FIG. 1. Computer 100 includes a central processor 110, a system memory 112 and a system bus 114 that couples various system components including the system memory 112 to the central processor unit 110. System bus 114 may be any of several types of bus structures including a memory bus or memory controller, a peripheral bus, and a local bus using any of a variety of bus architectures. The structure of system memory 112 is well known 25 to those skilled in the art and may include a basic input/output system (BIOS) stored in a read only memory (ROM) and one or more program modules such as operating systems, application programs and program data stored in random access memory (RAM).

WO 03/054756 PCT/IB02/05580 6 Computer 100 may also include a variety of interface units and drives for reading and writing data. In particular, computer 100 includes a hard disk interface 116 and a removable memory interface 120 respectively coupling a hard disk drive 118 and a removable memory drive 122 to system bus 114. Examples of removable memory drives include magnetic disk 5 drives and optical disk drives. The drives and their associated computer-readable media, such as a floppy disk 124 provide nonvolatile storage of computer readable instructions, data structures, program modules and other data for computer 100. A single hard disk drive 118 and a single removable memory drive 122 are shown for illustration purposes only and with the understanding that computer 100 may include several of such drives. Furthermore, computer 10 100 may include drives for interfacing with other types of computer readable media. A user may interact with computer 100 with a variety of input devices. Figure 1 shows a serial port interface 126 coupling a keyboard 128 and a pointing device 130 to system bus 114. Pointing device 128 may be implemented with a mouse, track ball, pen device, or similar device. Of course one or more other input devices (not shown) such as a joystick, game pad, satellite 15 dish, scanner, touch sensitive screen or the like may be connected to computer 100. Computer 100 may include additional interfaces for connecting devices to system bus 114. Figure 1 shows a universal serial bus (USB) interface 132 coupling a video or digital camera 134 to system bus 114. An IEEE 1394 interface 136 may be used to couple additional devices to computer 100. Furthermore, interface 136 may configured to operate with particular 20. manufacture interfaces such as FireWire developed by Apple Computer and i.Link developed by Sony. Input devices may also be coupled to system bus 114 through a parallel port, a game port, a PCI board or any other interface used to couple and input device to a computer. Computer 100 also includes a video adapter 140 coupling a display device 142 to system bus 114. Display device 142 may include a cathode ray tube (CRT), liquid crystal display 25 (LCD), field emission display (FED), plasma display or any other device that produces an image that is viewable by the user. Additional output devices, such as a printing device (not shown), may be connected to computer 100. Sound may be recorded and reproduced with a microphone 144 and a speaker 166. A sound card 148 may be used to couple microphone 144 and speaker 146 to system bus 114. One 30 skilled in the art will appreciate that the device connections shown in figure 1 are for illustration WO 03/054756 PCT/IB02/05580 7 purposes only and that several of the peripheral devices could be coupled to system bus 114 via alternative interfaces. For example, video camera 134 could be connected to IEEE 1394 interface 136 and pointing device 130 could be connected to USB interface 132. Computer 100 may operate in a networked environment using logical connections to one 5 or more remote computers or other devices, such as a server, a router, a network personal computer, a peer device or other common network node, a wireless telephone or wireless personal digital assistant. Computer 100 includes a network interface 150 that couples system bus 114 to a local area network (LAN) 152. Networking environments are commonplace in offices, enterprise-wide computer networks and home computer systems. 10 A wide area network (WAN) 154, such as the Internet, may also be accessed by computer 100. Figure 1 shows a modem unit 156 connected to serial port interface 126 and to WAN 154. Modem unit 156 may be located within or external to computer 100 and may be any type of conventional modem such as a cable modem or a satellite modem. LAN 152 may also be used to connect to WAN 154. Figure 1 shows a router 158 that may connect LAN 152 to WAN 15 154 in a conventional manner. It will be appreciated that the network connections shown are exemplary and other ways of establishing a communications link between the computers may be used. The existence of any of various well-known protocols, such as TCP/IP, Frame Relay, Ethernet, FTP, HTTP and the like, is presumed, and computer 100 may be operated in a client-server configuration to permit a 20 user to retrieve web pages from a web-based server. Furthermore, any of various conventional web browsers may be used to display and manipulate data on web pages. The operation of computer 100 may be controlled by a variety of different program modules. Examples of program modules are routines, programs, objects, components, data structures, etc., that perform particular tasks or implement particular abstract data types. The 25 present invention may also be practiced with other computer system configurations, including hand-held devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCS, minicomputers, mainframe computers, personal digital assistants and the like. Furthermore, the invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a WO 03/054756 PCT/IB02/05580 8 communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices. Embodiments of the present invention utilize business level spreadsheet software to optimize the availability of products or merchandise throughout the supply chain, and especially 5 at the distribution center and retail location. In one embodiment, Microsoft's® Excel® software is used, however, the present invention is not intended to be limited to that embodiment, and other spreadsheet software may be used. Embodiments of the present invention may be used to quantify the numerous factors that may impact customer/consumer service, and the available trade-offs, and simulate varying aspects of the supply chain that may have an appreciable impact 10 on the operational efficiency and effectiveness of product flow. These factors may include, but are not limited to, the network structure, operational tactics, product proliferation, demand variability, forecasting processes, and planning processes. Furthermore, the impact on operating cost and revenue may be quantified. To this end numerous factors, such as the price of the product, the inventory, product availability, product lead-time, and product waste may be taken 15 into account. With regard to cost, simulations may account for manufacturing, warehousing, transportation, and financing costs, as well as the cost of poor service, namely lost sales and lost margin. Illustrative embodiments of the present invention achieve an efficient and streamlined supply chain by implementing various strategies to link suppliers, retailers and consumers. 20. Embodiments of the present invention may use any combination of the strategies described below. The first supply chain optimization strategy is to maintain an efficient assortment of items in each store. This strategy attempts to optimize the productivity of store space and to tailor shelf plans to the store trading area. A second strategy is efficient promotion, which is 25 designed to maximize the efficiency of trade and consumer promotions and to reduce forward buys and supply chain inventory. A third strategy is efficient product introductions. The goal is to reduce the time to market for new products, and to increase successful new introductions. The fourth strategy is efficient replenishment. According to this strategy, the time and cost in the replenishment system are optimized, and overall customer service is improved. The four 30 strategies may provide timely, accurate, paperless information flow, with smooth, continual product flow matched to consumption.

WO 03/054756 PCT/IB02/05580 9 A retail supply chain may include independent demand and product flows that are connected through inventory buffers at distribution center(s) and store(s). The inventory buffers ensure that differences between demand and supply are matched over time. For example, for effective replenishment, the inventory buffers are optimized to meet demand and maximize sales 5 while reducing inventory levels and costs. SUPPLY CHAIN OPTIMIZATION PROCESS Referring to Fig. 2, a method of optimizing a supply chain according to an embodiment of the invention is shown. - The first step 200 is a preparation step, in which the business requirements, strategy and operational configuration may be plotted. The next step 202 is a 10 segmentation step in which the business may be divided into appropriate relevant areas, and the representative stock keeping units (SKUs) are selected for each area. In step 204, files that need to be loaded into the system are formatted. The formatting step may include formatting the files into a spreadsheet format. After the data is formatted, the existing replenishment parameters are modeled, in step 206, to create a baseline analysis. The baseline may then be then used to 15 compare with other simulations, where each simulation represents a model of a possible future replenishment scenario. In the analysis/extrapolation step 208 a numerical comparison of the modeled scenarios may be made. PREPARATION The preparation step 200 may include multiple smaller steps. In particular, the current 20 business requirements and practices should be collected. In order to obtain the maximum benefit from such collection, benefit opportunities may be identified, the collection should be done with an understanding of current supply chain procedures, and resources, contacts, and sponsors may be identified. Next, a strategy may be defined. For example, the goal may be to restructure the wider 25 logistical framework. The focus may be on replenishment or the configuration of an optimal supply chain node. Another factor for defining the strategy is whether the goal is to benefit from better distribution capacity and improved lead times. The focus of the analysis within the business should be identified, i.e., channel, outlet, departments, etc.

WO 03/054756 PCT/IB02/05580 10 The preparation step 200 also may involve determining the operational configuration. This may entail identifying the replenishment/forecasting method, which accurately simulates the current method. Next, it may be necessary to determine whether the replenishment/ forecasting method needs to be modified to meet operational needs. Another step in determining 5 the operational configuration is to define store groupings based upon specific criteria, e.g., turnover. Consideration may be given to each detailed level parameter and the manner these parameters should vary by location and product type over time. In one embodiment, the last part of the preparation step 200 is a planning step that entails a high-level plan for extracting the data. 10 SEGMENTATION Creating a simulation of a supply chain that accounts for and tracks every product can be extremely complex. In order to simplify the simulation, a critical segmentation of SKUs may be selected that are representative of the overall business in step 202. The aim is to group similar products and distinguish products having different attributes. As a result, the SKUs may be 15 grouped so that the number of SKUs tracked is manageable while focusing on the key supply chain drivers. Typical segment types include rate of sale, flow/seasonal, cube, season, and value. In one implementation of the present invention, fifty products would be considered a typical number of products to model. Four SKUs per segment would result in no more than 12 segments (4 SKUs x 2 seasons x 6 ROS = 48 products). 20 The representative SKUs may be selected from each of the key segments. The selections may be made to provide a simplified picture of the complete product range and all demand and supply drivers. The data collected is set forth in step 204. The data may be provided in a spreadsheet format, such as Excel® format. Referring to Fig. 3, a block diagram illustrating one implementation of the data-formatting step 204 is shown. Data from the production systems 25 and/or database 210 is extracted in step 212 and exported in step 214, preferably in a spreadsheet format, such as Excel® format for use in the supply chain optimization system of the present invention. Alternatively, statistically valid demand data may be generated synthetically and used in place of production data, in order to evaluate new and unseen demand patterns, or to assist where the production data is not of sufficient quality. Synthetic parameters may be created in WO 03/054756 PCT/IB02/05580 11 step 211, and then synthetic data can be generated from the synthetic parameters in step 213. Finally, the resulting synthetic data may be exported in step 215. DATA FORMATTING The data may include weekly sales, weekly store replenishment, weekly demand, 5 monthly stock snapshots, and/or distribution center receipts. After the data has been received, the data may be validated and formatted prior to input, in step 204 (Fig. 2). A generating inventory tool application (GIT) may be used to convert the base files. For a distribution center application, the GIT application creates a regional distribution center (RDC) file using the distribution center gins and base files. The system then takes the files created by the GIT 10 application and calculates demand from weekly sales and inventory. The regional distribution center values are saved in a generated demand file. The system may also create a daily demand file that is inserted over the sales figures in the weekly demand file, which then becomes the Demand Added screen that may be used to create the model. Step 206 is the modeling of the existing replenishment parameters. By analyzing the 15 baseline replenishment strategy, future strategies may be performance modeled against the actual achieved results. The analysis provides detailed graphical and numeric output delineating, for example, the actual sales, transportation movement; inventory levels, and missed sales opportunities. The foregoing data may be reviewed at many levels, e.g., store level, store group level, or across the entire business. 20 After the baseline has been established, possible future flow paths and replenishment strategies may be modeled. The scenarios may be plotted to match very detailed replenishment strategies. The strategies may focus on replenishment options and include replenishment methods, desired service levels, the number of days of inventory cover, minimum presentation stock, and the number of replenishment days. Alternatively, the strategies may focus on supply 25 chain options including minimum store order quantity, minimum supplier order quantity, supplier to distribution center lead time, distribution center to store lead time, and distribution center safety stock. The detailed graphical and numeric output may be compared to the output for the baseline case.

WO 03/054756 PCT/IB02/05580 12 MODELING The last step is the analysis/extrapolation step 208. In step 208, the modeled product flow and replenishment criteria may be ranked against a quantifiable performance index (PI). The PI quantifies the cost/service tradeoff associated with various modeled scenarios. In one 5 embodiment, the PI is created using the Gross Profit After Distribution Expense (GPADE). The GPADE equals the sales revenue minus the landed cost of goods, the cost of store delivery, the cost of store operations, the cost of store inventory, the cost of the distribution center operations, and the cost of distribution center inventory. The resulting figure is used to produce a PI percentage using the following formula: 10 Simulation GPADE/Baseline GPADE = P.I. Index (%). In order to assess an optimal trade-off point, a cost-based analysis of all the factors in the PI may be made to ensure the most profitable decisions are taken. As a result, profit-making decisions are made which will remain robust under a variety of conditions. Referring to Fig. 4, a chart showing low and high profit scenarios is illustrated. Analysis of these scenarios shows that 15 if the SKU sold at a low profit, then the margin made on the product may be made by employing a low-inventory replenishment strategy, and hence, a lower service level. If the SKU sold for a higher profit then the more profitable decision for the business is to ensure a higher service level by making a larger investment in inventory. After the PI results have been obtained at the SKU level, benefits to the business may be 20 extrapolated to segment and business unit level. In one embodiment, the data for the selected representative SKUs within each segment is collated at a segment level, which in turn is collated at a business unit level. The basic workflow of the supply chain optimization system of certain embodiments of the present invention starts by calculating demand from the weekly sales and inventory. New 25 demand is then loaded with current store groups. Next the base case analysis is performed. The base case analysis yields the baseline for each representative SKU, and the results are recorded, for example in PowerPoint ®. The current supply chain is also plotted.

WO 03/054756 PCT/IB02/05580 13 Simultaneously, base case statistics from the base case analysis may be used to define supply availability characteristics. The forecast profile is then defined. From the forecast profile, the replenish characteristics are defined. The simulation is then run and the inventory and availability are plotted. The process, starting from defining the replenishment characteristics, 5 may be repeated as necessary to determine the optimum cover level for the supply chain. When the optimum cover level is found, push pull analysis may then be performed, and the results may be archived. INTERFACE SCREENS Figure 5 illustrates an exemplary main menu screen 300 that may be used with 10 embodiments of the present invention. A configuration box 301 on the main menu 300 contains a setup screen button 302, which allows the user to select desired screen layouts from the main menu 300. A setup names button 304, also in the configuration box 301, brings up screen 306, shown in Fig. 6, where the user may control the names of the groups of stores, usually during initial setup. The configuration box 301 also contains a setup store groups button 308. By 15 selecting button 308, the user is presented with screen 310, shown in Fig. 7. From screen 310, the user may define store groupings based on specified criteria, such as sales volume. After completing the configuration of the supply chain via the configurations box 301, a user next proceeds to the simulate demand box 311 in the main menu 300, shown in Fig. 5. The simulate demand box 311 has three buttons, a define demand characteristics button 312, a 20 generate synthetic demand button 314, and a load new demand with current store groups button 316. Selection of the define demand characteristics button 312 brings up demand characteristics screen 400, shown in Fig. 8, through which the user may enter data defining the product demand in the supply chain being analyzed. The data here may include the total number 25 of stores, weekly and seasonal demand pattern data, product launch/ termination data, promotional activity data, and base demand data. The button 314 generates synthetic demand, as opposed to utilizing existing demand data. Synthetic demand may be beneficial in testing a previously configured supply chain.

WO 03/054756 PCT/IB02/05580 14 Button 316 allows the user to load new demand data into the model, using the current store groups that were set up in through the screen 310. The data files that are used to load the demand into the model should contain exactly the same number set up in the store group screen 310, so that each store is aligned with the correct store data, and so that the data may be sorted 5 into groups. The button 316 allows the user to load a previously saved set of store groups from a text file. The define supply chain and replenishment characteristics box 317 on the main menu 300, shown in Fig. 5, has three buttons; a define replenishment parameters button 318, a define profile forecast characteristics button 320, and a define supply availability characteristics button 10 322. By selecting the define replenishment characteristics button 318, a user is presented with the screen 324, shown in Fig. 10. From the replenishment characteristics screen 324 a user may select which facilities to model in the distribution network, how many facilities in the network, operations within each facility, and how the product will be moved or replenished to each 15 facility. The replenishment characteristics screen 324 contains a set parameters button 326, which, when selected by the user, presents the further replenishment characteristics menu 330, shown in Fig. 11. The replenishment characteristics menu 330 allows the user to input information, such as lead times, pack size quantities, and transportation modes. The supplier/distribution pack size is the number of units received for a particular SKU for one 20 order. The regional distribution center lead-time is the time taken between an order being placed with the supplier and the distribution center receiving the stock. Regional distribution center review time is the time for reviewing the inventory levels at the regional distribution center. The user may further adjust the change over time between push and pull strategies. The user may further allow stock levels to be run down near the end of an on-sale period. 25 The replenishment characteristics menu 330 contains a replenishment methods button 336, a replenishment days button 338, a forecasting button 340, an initialization button 332, a set allocations button 334, and a box 340 to limit supplier availability. The replenishment days button 338 on the replenishment characteristics menu 330 brings up the replenishment days screen 460, shown in Fig. 19. The user may specify how frequently WO 03/054756 PCT/IB02/05580 15 and on which days various store groups should be replenished through the replenishment days screen 460. One skilled in the art will appreciate that numerous alternative interface screens may be utilized to allow users to input and view data. 5 PARAMETERS One implementation of the supply chain optimization system and method according to the present invention is to provide a retail allocation and replenishment simulation model. There are a number of parameters that control the synthetic demand, such as: Total Stores; Average Daily Demand for the chain; SD Daily Demand for the chain; Seasonal Profile; Launch Week 10 for the chain; Launch Period for the chain; Termination Week for the chain; Tail-off Period for the chain; Number of Days; Earliest Start Week for the chain; Earliest End Week for the chain; Minimum Duration for the chain; Maximum Duration for the chain; Minimum Interval for the chain; Maximum Interval for the chain; Number of Stores for each store group; Store Size Relationship for each store group; Average Daily Demand for each store group; SD Daily 15 Demand for each store group; Weekly Profile for each store group; Launch Demand (Mean) for each store group; Launch Demand (SD) for each store group; Relative Uplift (Mean) for each store group; and Relative Uplift (SD) for each store group. In one particular implementation of the invention, three types of synthetic demand pattern may be used, namely (1) conseassional, (2) seasonal, and (3) promotional, through the 20 demand characteristics screen 400, shown in Fig. 8. The average daily demand for each store may be calculated as a random number from a normal distribution with a mean equal to the average demand for all stores in the same store group and with a standard deviation as defined for the store group. If the average demand is calculated as being less than zero then the model equates it to zero. 25 The promotional weeks may be set using a sampling process and an incrementation method using the sum of two uniform distributions constructed as Int (earliest start week for the promotion plus Rnd 0 multiplied by (earliest end week for the promotion plus 0.99999 minus earliest start week for the promotion)) plus Int(minimum duration for the promotion plus Rnd0 WO 03/054756 PCT/IB02/05580 16 multiplied by (maximum duration for the promotion plus 0.99999 minus minimum duration for the promotion)) Further promotional weeks are then further increment by the addition of another uniform distribution or the form Int(minimum interval for the promotion plus Rnd0 multiplied by 5 (maximum interval for the promotion plus 0.99999 minus minimum interval for the promotion)) and proceeding to increment with the above distributions until a full year has elapsed. The calculation of extra demand is dependent on where in the period the model is in the simulation. Thus, if the week in the simulation under consideration is between the start week and the end week, then extra demand is calculated as the average daily demand multiplied by the 10 store group daily demand from the weekly profile multiplied by the seasonal profile for the current week If the day in the simulation under consideration is the first day in the start week, then the initial demand for each store is calculated as a random number generated from a normal distribution whose mean is defined as the average launch demand for the store group to which 15 the store belongs and launch standard deviation defined for the store group. If the week in the simulation is within the tail-off period, then the extra demand is calculated as average daily demand multiplied by the store group daily demand from the weekly profile multiplied by the seasonal profile for the current week multiplied by the difference between the end week and the current week, divided by the length of the tail-off period. 20 If the week in the simulation is between the launch week and the launch period then the extra demand is calculated as (initial demand multiplied by the store groups daily profile for the current day multiplied by the (Launch week plus the tail off period minus the current week) plus (average daily demand multiplied by the store group daily demand from the weekly profile multiplied by the seasonal profile for the current week) multiplied by (current week minus the 25 launch week)) divided by the length of the launch period. If the week in the simulation is a promotional week then the extra demand is calculated as the current extra demand previously calculated multiplied by (one plus (a random value from a normal distribution whose mean is defined as the average launch demand for the store group WO 03/054756 PCT/IB02/05580 17 that the store belongs and launch standard deviation defined for the store group multiplied by store group demand)). The demand for the store on each current day is given as a sample from a Poisson distribution whose mean is given by the relevant extra demand. 5 PARAMETERS USED IN THE IMPLEMENTATION OF THE MODEL The following parameters are used in implementing a model in accordance with an illustrative embodiment of the invention, via the demand characteristic screen 400, shown in Fig. 8: Average Daily Demand for the chain; SD Daily Demand for the chain; Seasonal Profile for the chain; Launch Week for the chain; Launch Period for the chain; Termination Week for 10 the chain; Tail-off Period for the chain; Number of Days for the chain; Earliest Start Week for the chain; Earliest End Week for the chain; Minimum Duration for the chain; Maximum Duration for the chain; Minimum Interval for the chain; Maximum Interval for the chain; Store Size Relationship for each store group; Weekly Profile for the smallest store group, Mon, Tues, Wed, Thurs, Fri., and Sat; Launch Demand (Mean) for the smallest store group; Launch 15 Demand (SD) for the smallest store group; Relative Uplift (Mean) for the smallest store group; and Relative Uplift (SD) for the smallest store group. The values that are updated from the "store groups" screen include the number of Stores in each store group. The values are linked to the values: - The Average Daily Demand for each store group; 20 - The SD Daily Demand for the each store group; - The Weekly Profile for the smallest store group, Sun profile is calculated as 100 sum of the entered profiles. The weeks profile for the rest of the store groups are given to be the same as the smallest store group; - The Launch Demand (Mean) for the rest of the store groups is given to be the same 25 as the smallest store group; - The Launch Demand (SD) for the rest of the store groups are given to be the same as the smallest store group; WO 03/054756 PCT/IB02/05580 18 - The Relative Uplift (Mean) for the rest of the store groups are given to be the same as the smallest store group; and - The Relative Uplift (SD) for the rest of the store groups is given to be the same as the smallest store group. 5 The values are read into the model by the demand generation procedures when the button 506 is selected in main menu 300. The Seasonal profile value entered (range 1 to 20) on the demand characteristics screen 400 may be used to determine the seasonal profile week used in the demand generation, which is held on the Profiles screen 404, shown in Fig. 9. 10 ACTUAL DEMAND Actual demand is based on actual sales and inventory. Actual demand generation is defined from the main menu 300 (Fig. 5) using button 316. The model may use two forms of data to generate actual demand based on either weekly sales and inventory data or daily sales and inventory data. 15 WEEKLY DEMAND AND INVENTORY DATA The following sets forth the definition and impact of data on the model. First is the weekly sales and inventory data. The model reads in the data from an external file which may be in a spreadsheet format and which contains the two screens (not shown) upon selection of button 502 in the main menu 300, shown in Fig. 5; a sales screen containing the weeks sales for each 20 store in the chain and a stock screen containing the weeks stock at each store in the chain. While the weekly data is read into the model, the data is checked for negative values in sales and sets them to zero. Negative stock is treated in a slightly different manner. The modulus of the largest negative value in all the stock data is taken and then added to all the stock data in the chain. When the user selects the calculate demand from weekly sales and inventory button 502, 25 the user is instructed to select a data file to be loaded into the system. The system takes the data and uses it as an average across the season and artificially estimates the breakdown of sales, demand and stock into individual days, and sales these in SKU files.

WO 03/054756 PCT/IB02/05580 19 Generation of the demand for the data read into the model involves the model manipulating the data before the demand may be calculated. The model takes the data and converts the read-in weekly data into daily data as follows. For each week and store the model generates the daily sales data by taking a temporary value generated by integer value of (1 + 5 Random number * (100 + 0.99999 - 1)). This value is generated each time until the demand reaches the value of the week's sales for the store. Each time the value is generated it is checked against a daily distribution. The daily distribution on day i is calculated as the given daily distribution on day i + daily distribution on day (i - 1), (for i 1). If the value is between 1 to the daily distribution on 10 day 1, then a sale is generated on day 1 of the week under consideration. If the value is between (daily distribution on day 1) + 1 to daily distribution on day 2, then a sale is generated on day 2 of the week under consideration. If the value is between (daily distribution on day 2) + 1 to daily distribution on day 3 then a sale is generated on day 3 of the week under consideration. If the value is between (daily distribution on day 3) + 1 to daily distribution on day 4 then a sale is 15 generated on day 4 of the week under consideration. If the value is between (daily distribution on day 4) + 1 to daily distribution on day 5 then a sale is generated on day 5 of the week under consideration. If the value is between (daily distribution on day 5) + 1 to daily distribution on day 6 then a sale is generated on day 6 of the week under consideration. If the value is between (daily distribution on day 5) + I to daily distribution on day 6 then a sale is generated on day 6 20 of the week under consideration. If the value is between (daily distribution on day 6) + 1 to daily distribution on day 7 then a sale is generated on day 7 of the week under consideration. Next the model converts stock data to daily by first considering the weekly stock and sales data. The model checks that the stock at the end of each week is equal to the stock at the end of the week before subtracting the sales for the week. If this is not true then the model 25 assumes the store has had a delivery during the week and works out the amount delivered and sets the delivery day as the value generated from the calculation Int (1 + Random number (between 0 and 1) * (7 + 0.99999 - 1)). The stock for sale at the beginning of the first week of the data is assumed to be the week's stock plus the week's sales. The stock at the beginning of each of the following week is 30 the stock level at the end of the previous week.

WO 03/054756 PCT/IB02/05580 20 The daily stock is the stock on the first day of the week minus the daily sales generated earlier. If this causes a stock out, i.e., the sales are greater than the stock, then the model pulls forward the delivery day and sets the daily stock to the stock the day before, minus the daily sales, plus the amount delivered (calculated earlier from the weekly data). 5 The model then generates new files in the data's external file and writes the daily sales and daily stock to these files after the data is sorted into store groups defined in the store groups' files. The model orders the stores in the order given in the screen 306, shown in Fig. 6, with the first group being the smallest stores. Now the data is in a format to generate demand. DAILY SALES AND INVENTORY DATA & DEMAND GENERATION 10 The model reads in the daily data and formats the negative data as for weekly, and then sorts the data into store groups order upon selection of button 504 on the main menu 300, shown in Fig. 5. The model proceeds to generate daily demand. If daily stock for a store on any of the days in the simulation is equal to zero then the model finds the store group frequency of sales. This frequency is then normalized such that the model may generate a random number that may 15 intersect the cumulative frequency of the store group sales, using that value as the expected sales values. The cumulative frequency is the sum of all stores that have sales less than or equal to a predetermined number. If the model generates "extra" sales using this method then the model uses these extra sales, otherwise the model uses original sales data for the demand. The following is an example, where for day three of the simulation, store I has no stock. 20 If store 1 is a member of "store group large" which contains: Store Sales on day 3 1 2 2 3 3 4 4 3 5 0 Then the maximum sales for the large store group is 4 units.

WO 03/054756 PCT/IB02/05580 21 The frequencies of the sales units in the store group are: - Sales of 2 units has a frequency of 1 - Sales of 3 units has a frequency of 2 - Sales of 4 units has a frequency of I 5 - Sales of 0 units has a frequency of I - 5 stores in the group had sales. The conversion to cumulative frequencies [0,1] is: - Cumulative frequency of no sales = 1 / 5 - Cumulative frequency of 1 sale = 1 / 5 10 - Cumulative frequency of 2 sales = 1 / 5 + 1 / 5 = 2/5 - Cumulative frequency of 3 sales = 2 / 5 + 2 / 5 = 4/5 - Cumulative frequency of 4 sales = 4 / 5 + 1 / 5 = 1 The model generates a random number of 0.8. The model finds that the number of expected sales is 3 sales and, as this is greater than the actual sales generated, the demand for 15 store I is set to 3 units. The daily distributions are entered in the Demand characteristics screen 400. The values are placed on the screen as follows: - Given daily distribution on day 1; - Given daily distribution on day 2; 20 - Given daily distribution on day 3; - Given daily distribution on day 4; - Given daily distribution on day 5; and - Given daily distribution on day 6. SUPPLY CHAIN CONFIGURATION: STORE GROUPS 25 We turn now to the store groups of the supply chain configuration. The stores as a whole may be split up into smaller groups of stores that are similar, or share similar properties. This allows similar stores to be treated the same, yet still allowing variation between stores. Some examples of the properties used to group stores may be size, location or type of store.

WO 03/054756 PCT/IB02/05580 22 The replenishment model allows many parameters to be set by store group. This results in greater flexibility, and allows different types of stores to be treated differently. The following parameters, all used in replenishment algorithms, may be set differently for different store groups: service level, inventory selling days, presentation stock, minimum stock, maximum 5 stock, weeks of cover, and allocation quantity. For example, the Min-Max replenishment method requires that the minimum (and maximum) stock levels for stores be defined. Clearly, different size of stores may use different parameters for the minimum stock level. In a smaller store the minimum allowable stock may be 1 unit, while in a larger store the minimum should be 3 units. By allowing parameters to be set by store groups, greater flexibility may be achieved. 10 The store groups are controlled through the Store Group screen 310, shown in Fig. 7. From the main menu 300, this is accessed via the Set-up Store Group button 308. The store group names are displayed on the Store Group screen 310, and are referenced from the Names screen 306, shown in Fig. 6, and should only be changed on the Names screen 306. The Store Group screen 310 should contain the actual store numbers of the stores being 15 analyzed. The store numbers may be unique. If synthetic data is being used, then any unique set of store numbers will suffice, as long as the total in each store group is correct. The first store number in each store group may be placed in row 2. The data files used to load demand into the model should contain exactly the same store numbers that appear in the Store Group screen 310. This is necessary to allow each store to be aligned with the correct store 20 data, and to allow the data to be sorted into store groups. The stock allocation routines, discussed below, distribute any available stock to the largest stores first, and use the store groups to determine which stores to allocate first. Store group 1, the stores held in column A of the Store Group screen 310, is defined as the store group which contains the smallest stores. Store Group 6, the stores held in column F of the Store 25 Group screen 310, is defined as the store group containing the largest stores. In the example illustrated in Fig. 7, there are only four sizes of stores, so column D contains the largest stores. Within each store group, the store held in row 2 should be thought of as the smallest store. This store generally receives stock last in this store group. The bottom store in each store group should be thought of as the largest store, or the store that will receive stock first.

WO 03/054756 PCT/IB02/05580 23 The Save Demand & Current Store Groups button 508, in the main menu 300, allows the user to save data after it has been entered. Similarly, the Recover Saved Demand & Current Store Groups button 510 prompts the user to select a file from which demand, sales, and stock data may be loaded into the system. Selection of button 510 loads the relevant demand, sales and 5 stock information into active memory of the computer system 100. Every time the system of the present invention is started the data need to be reloaded using the button 510, or new data need to be entered by the user. Store groups may also be used in the calculation of any extra demand. Once the store numbers have been inserted the Update Screen & Exit, button 319 calculates the number of 10 stores in each store group and writes this information to the Store Group screen 310. It also writes the size of each store group to the Demand Characteristics screen 400. The Save Store Groups button 313, shown in Fig. 7, allows the user to save the current store group settings. The store groups are saved as a text file. The Load Store Groups button 315 allows the user to load a previously saved set of store groups from a text file. The model 15 prompts the user with standard loading dialogue. All procedures are held in the Store Groups module. SUPPLY CHAIN CONFIGURATION: FLOW OPTION The flow option of the supply chain configuration in accordance with an embodiment of the invention will now be described. There are a number of different flow options, which define 20 the method used to allocate stock to both the distribution center and the stores. The current flow options are controlled on the Replenishment Characteristics screen 324, shown in Fig. 10, which is accessed from the main menu 300, shown in Fig. 5, via the Define Replenishment Characteristics button 318. A set parameters button 326 allows the user to control which method is used. The six flow options available are single push, push in waves, push - controlled push, 25 push pull, continuous pull, and pull - cross docked. The selected flow option number is written to the Supply Characteristics screen 700, shown in Fig. 25. The flow option name is referenced from the Setup Names screen 306 using the flow option number. A flag is used to determine whether cross docking is enabled and is calculated in the Supply Characteristics screen 700. Another flag is used to determine whether a WO 03/054756 PCT/IB02/05580 24 push has been selected, and is calculated in the Supply Characteristics screen 700. The flags are read into the model in the Replenishment Module. SINGLE PUSH Turning now to the single push option, this flow option may be used when all stock 5 purchased is to be dispatched to the stores. The amount of stock available to be allocated is controlled via the Initial Purchase Quantity text box 331 on the Replenishment Characteristics screen 330. This value is output to the Supply Characteristics screen 700. The distribution of the available stock is controlled 'ia the Allocations screen 408, shown in Fig. 13, which is accessed using the Set Allocations button 334 on the Replenishment Characteristics menu 330. The 10 timing of the push, and the quantity pushed to each store are predetermined. The distribution center is never replenished; the Initial Purchase Quantity is the only stock available. PUSH IN WAVES The next option is push in waves. This flow option allows the available distribution center stock to be dispatched to the stores in a series of pushes. The amount of stock available 15 for allocation is controlled via the Initial Purchase Quantity text box 331. This value is output to the Supply Characteristics screen 700. The distribution of the available stock is controlled via the Allocations screen 408, which is accessed via the Set Allocations button 334 on the Replenishment Characteristics screen 330. Again, the timing of each push, and the quantity pushed to each store, are predetermined, before the simulation begins. Furthermore, the 20 distribution center is never replenished; the Initial Purchase Quantity is the only stock available. PUSH - CONTROLLED PUSH The next option is push - controlled push, which allows the available distribution center stock to be dispatched to the stores in a series of pushes. The timing of each push is fixed, but the quantities pushed to each store are not predetermined. The initial push quantity is, however, 25 fixed but subsequent push sizes are determined by the forecasted sales, in each store. The total amount of stock available to be allocated is controlled via the Initial Purchase Quantity text box 331. This value is output to the Supply Characteristics screen 700. The distribution of the available stock is controlled via the Allocations screen 408, which is accessed WO 03/054756 PCT/IB02/05580 25 via the Set Allocations button 334 on the Replenishment Characteristics screen 330. For Push Controlled Push, the amount allocated to each store is unimportant. The actual quantity dispatched to a store will be dependent on the forecasted sales. When defining the size of the pushes, the important quantity is the total amount being distributed to the stores, for a particular 5 push. When the allocation quantities are finalized, the fraction of the total available stock that is pushed is calculated for each push. The values are written to the Supply Characteristics screen 700, where the fractions are converted to the percentages. The percentages pushed are used to calculate the actual amount allocated in each push. These values are used in the calculation of 10 the controlled push. The following calculation explains the calculation of the Controlled Push. In the explanation below, the store forecasts are the number of sales forecast between the current time and the time of the next push. The total forecast is the sum of all the store forecasts. The total allocation quantity is the total amount to be allocated to the stores in this push. The amount that 15 is actually allocated may exceed the total allocation quantity, if the forecasts are large. However, the total that is available for allocation over all the pushes is still fixed by the Initial Purchase Quantity. The controlled push may be calculated as follows: Forecast Stock = Stock on Hand - Store Forecast Store Order Quantity = (Total Allocation(Push) + Total Forecast) x Store Forecast 20 Total Forecast - Forecast Stock The store order quantity is then rounded up to the nearest store minimum order quantity multiple.

WO 03/054756 PCT/IBO2/05580 26 Example: Parameter Value Stock on Hand 2 Store Forecast 3 Total Forecast 9 Total Allocation 6 Store Order Quantity = (Total Allocation(Push) + Total Forecast) x Store Forecast Forecast Stoc Total Forecast (6 + 9 ) x 3( 2 3 ) = - (2 -3) 9 =5+1 =6 units This procedure may be repeated for each store, or until all the available stock has been 5 used up. PUSH PULL The next option is the push pull. For this method, an initial allocation, the push, is made, and the items are dispatched to the stores. Once the initial allocation has been sent, the stores switch to a replenishment method, and pull stock from the distribution center. The appropriate 10 replenishment algorithm determines the quantities pulled by the stores. The timing of the initial push, and the quantity pushed to each store are predetermined. The Initial Purchase Quantity text box 331 on the Replenishment Characteristics screen 330 controls the amount of stock available to be allocated. This value is output to the Supply Characteristics screen 700. The distribution of the available stock is controlled via the 15 Allocations screen 408, which is accessed via the Set Allocations button 334 on the Replenishment Characteristics screen 330. Once the initial allocation has been made, a changeover time between push and pull may be defined. This is controlled via the Replenishment Characteristics screen 330, the week of the changeover takes place in is written to the Supply Char screen 700.

WO 03/054756 PCT/IB02/05580 27 After the changeover has been made from push to pull, the replenishment algorithm determines the order quantities for each store. Stock continues to be pulled from the distribution center until the end of the simulation period, or until there is no stock remaining. The distribution center is never replenished; the Initial Purchase Quantity is the only stock available. 5 CONTINUOUS PULL Continuous pull is the next flow option, and uses replenishment to control both the stock level in the store and in the distribution center. For the distribution center, the replenishment algorithms control when, and how much stock is ordered from the supplier. For the store, the replenishment algorithms control when, and how much stock is ordered from the distribution 10 center. The different replenishment methods control when, and how much stock is ordered from the supplier, and when stock is ordered from the distribution center. The stock is initialized at a particular value, and the chosen replenishment algorithm runs. By using a continuous pull, both the distribution center and the stores are replenished. PULL CROSS DOCKED 15 The next flow option is pull cross docked, which uses replenishment algorithms to control the stock in the stores, but without any distribution center stock. Stock is delivered from the supplier and is immediately dispatched to the stores. The supplier always satisfies the stores order quantities, unless the Supplier Availability options are selected, whereby the amount delivered by the supplier may be restricted. The replenishment algorithms generate the store 20 order quantities. SUPPLY CHAIN CONFIGURATION: LEAD-TIMES The next topic is lead-times of the supply chain configuration. The lead-time parameters are defined on the Replenishment Characteristics screen 330 and are in units of days. The model uses two lead-time parameters, namely, distribution center lead-time (range 1 to 197), and store 25 lead-time (range I to 13). The lead-time defines the number of days between when an order is issued and when the resulting inventory is available either at the distribution center or store. For example in the scenario where store lead-time is 1 day, if an order is issued at the end of trading on Tuesday, WO 03/054756 PCT/IB02/05580 28 the stock will arrive at the store for the beginning of trading on Wednesday. The following is an example: Lead-time is 1 day; At the beginning of Day 10: The stock at store A is 2 units (i.e. the stock at the end of 5 day 9 assuming no deliveries); The demand for day 10 is 2 units; End of day 10: the closing stock is 0 units; End of day 10: assume the replenishment system generates an order for 3 units; At the beginning of day 11: the stock is 0 units (stock at the close of day 10) + 3 units 10 (stock ordered at the end of day 10 which is delivered for the start of day 11 since the delivery lead-time is 1 day) To implement the foregoing in the model, the lead-times are entered in the Replenishment Characteristics screen 330. The values entered are placed on the Supply Characteristics screen 700. Distribution center lead-time is used for the calculation of 15 forecasting and replenishment times on the Supply Characteristics screen 700. The values are read into the model according to the replenishment. INVENTORY INITIALIZATION: STORE INVENTORY At the start of each simulation run, the stock level in each store is initialized to a pre determined level. The initial stock level to be set in each store is shown on the Supply 20 Characteristics screen 700 (Fig. 25). This level may vary by store group. The initial stock levels are dependent on the chosen replenishment method. For all replenishment methods, excluding the dynamic method, the initial stock level is typically set to the minimum stock level. There are two methods of defining the minimum stock level. This may be either a specific value, e.g. 3 units, or a number of day's cover, e.g. 10 days worth of sales. The method chosen to define the 25 minimum stock level is set in the Replenishment Methods screen 420, shown in Fig. 14a, for distribution centers, and screen 421 shown in Fig. 14b, for stores and is written to the Supply WO 03/054756 PCT/IB02/05580 29 Characteristics screen 700. Screens 420 and 421 are both accessed by selecting the replenishment methods button 336 on screen 330. If the minimum stock level is defined as a specific value, then this parameter is used directly as the initial stock level. This minimum parameter is set in the Replenishment Methods 5 screen 420. If the minimum value is defined as a number of day's cover, then the initial stock level is set as the number of days cover x average daily store demand. This number is rounded up to the nearest integer. The number of days cover is also set in the Replenishment Methods screen 420 and is written to the Supply Characteristics screen 700. From the Replenishment Methods screen 421, the user may activate the forecast methods 10 screen 430, shown in Fig. 16, by selecting the forecasting button 424. By selecting the set parameters button 432, the forecast parameters screen 440, shown in Fig. 17, is presented to the user. If the user selects the set profiles button 444 in the forecast parameters screen 440, the winters seasonal profiles screen 450, shown in Fig. 18, is presented. In the case of the dynamic method, the initial stock level is set as the presentation stock. 15 This parameter may be set on the Replenishment Methods screens 421, when the dynamic method is chosen for store replenishment and is written to the Supply Characteristics screen 700. OVER-RIDING INITIAL INVENTORY The model may provide the ability to override the initial stock levels calculated above. By checking the Use Initial Stock Levels check box, on the Replenishment Characteristics 20 screen the model will use the actual initial stock levels from the data set, with some caveats. If the stock at the beginning of the simulation period is less than the initial target stock level, calculated above, then the stock is set to be the target level, and any stock already in the store supply chain is removed. When data is loaded into the model, from an external data file, the "actual" initial stock data is also calculated, by adding the stock level at the end of the first 25 day and the sales during the first day.

WO 03/054756 PCT/IB02/05580 30 If the actual initial stock is greater than the target stock level, then the stock at the store is set to the actual initial stock level, i.e. the largest of the tow values. Otherwise, the stock remains at the target level. DISTRIBUTION CENTER INVENTORY 5 At the start of each simulation run, the stock level in the distribution center may be initialized to a pre-determined level. The initial stock level to be set in the distribution center is shown on the Supply Characteristics screen 700. The distribution center stock level is initialized as the maximum distribution center stock. Initial Stock = Maximum Stock = Minimum + Supplier MinimumOrderQuantity / 2 O10 FORECAST INITIALIZATION The following is an explanation of forecast initialization. One parameter controls the entire forecast initialization procedure. This parameter may be set on the Forecast Parameters screen 440, shown in Fig. 17, accessed through button 423 in screen 421, shown in Fig. 14b. The parameter labeled 'Data Points to initialize with' controls the number of weeks used to calculate 15 the initial values for the forecasting. The parameter is written to the Supply Characteristics screen 700. Each store is initialized separately, in the following explanation; it is assumed that the initialization is for an individual store. At least five methods may be used for store forecasting in accordance with illustrative embodiments of the invention. The level of the series estimates the non-seasonal, slowly 20 changing process of the time series. This data feature represents a flat value for the data when noise, trend and seasonality effects are excluded from the data. The trend of the series reflects the rate of change of the series from one point to another. The five illustrative methods are: moving average having level parameters, simple exponential smoothing having level parameters, Holts method having level and trend parameters, Generic 25 method having forecast parameters, and external having forecast parameters. For the moving average, the level is initialized by calculating the average weekly sales in the store, over the number of weeks defined by the initialization parameter.

WO 03/054756 PCT/IB02/05580 31 Level = Total Store Sales (Period) Initialise Weeks For simple exponential smoothing, the level is initialized by calculating the average weekly sales in the store, over the number of weeks defined, by the initialization parameter. Level = Total Store Sales (Period) Initialise Weeks 5 Holts method is a more advanced forecasting method, which requires the initialization of both the level and trend parameters. The level initialized by calculating the average weekly sales in the store, over the number of weeks defined by the initialization parameter. Level = Total Store Sales (Period) Initialise Weeks The trend is initialized using the following formula: 10 Trend = Store Sales (Last Week) - Store Sales(LastWeek - Initialise Weeks) InitialiseWeeks - 1 The Generic method uses only sales from the last week to initialize the forecast. The forecast is initialized as the demand from last week. Level = Total_ Store _ Sales (Last Week) For the external forecast, no initialization, as such, is performed; instead the entire 15 forecast is read in from an external file. To initialize the total forecast, one parameter controls the forecast initialization procedure. This parameter may be set on the Forecast Parameters screen 440, shown in Fig. 17. The parameter labeled 'Data Points to initialize with' controls the number of weeks used to calculate the initial values for the forecasting. The parameter is written to the Supply WO 03/054756 PCT/IB02/05580 32 Characteristics screen 700. The same parameter is used for both the stores forecast initialization, and the total forecast initialization. At least six methods may be used for store forecasting in accordance with illustrative embodiments of the invention. The methods used, and the parameters that need to be initialized 5 are as follows: moving average having level parameters, simple exponential smoothing having level parameters, Holts method having level and trend parameters, Winters Additive having level and trend parameters, Winters Multiplicative having level and trend parameters and Generic method having forecast parameters. For the Moving Average the level is initialized by calculating the average weekly sales 10 in all stores, over the number of weeks defined by the initialization parameter. Level = Total StoreSales (Period) Initialise Weeks For simple exponential smoothing, the level is initialized by calculating the average weekly sales in all stores, over the number of weeks defined by the initialization parameter. Level = Total StoreSales (Period) Initialise Weeks 15 Holts Exponential smoothing method is a more advanced forecasting method, which requires the initialization of both the level and trend parameters. The level initialized by calculating the average weekly sales in all stores, over the number of weeks defined by the Initialization parameter. Level = Total StoreSales (Period) Initialise Weeks 20 The trend is initialized using the following formula: Trend = Total Sales (Last Week) - Total Sales(LastWeek - Initialise Weeks) InitialiseWeeks - 1 WO 03/054756 PCT/IB02/05580 33 Winters Exponential Smoothing (Additive) Method is a more advanced forecasting method, which requires the initialization of both the level and trend parameters. The level initialized by calculating the average seasonalized weekly sales in all stores, over the number of weeks defined by the initialization parameter. 5 Level = Total Sales- Total Seasonal Index 5 Level = Initialise Weeks The trend is initialized using the following formula: [(Total Sales (Last Week) - Seasonal Index (Last Week)) Trend = - (Total Sales (Last Week -Initialise Weeks) - Seasonal Index (Last Week - Initialise Weeks)j Initialise Weeks - 1 The seasonal index, or profile, is used as a guide to the seasonality in the data. It is 10 usually scaled, aggregated set of the previous years sales. The index is held on the winters screen 640, shown in Fig. 24. This data is read into the model in the Replenishment Module. Winters Exponential Smoothing (Multiplicative) Method is a more advanced forecasting method, which requires the initialization of both the level and trend parameters. The level initialized by calculating the average seasonalized weekly sales in all stores, over the number of 15. weeks defined by the initialization parameter, i.e.: Level Total Sales Level = Initialise Weeks x Total Seasonal Index' The total seasonal index is the product of the seasonal indices of the weeks under consideration. Total Seasonal Index = rI Seasonal Index 20 The trend is calculated using the seasonalized sales values, and is initialized using the following formula: WO 03/054756 PCT/IB02/05580 34 Total Sales (Last Weekl ISeasonal Index (Last Week)) (Total Sales (Last Week -Initialise Weeks)W/ (Total Sales (Last Week -nitialise W Seasonal Index (Last Week - Initialise Weeks Trend = Initialise Weeks -1 The seasonal index, or profile, is used as a guide to the seasonality in the data. It is usually a scaled, aggregated set of the previous years sales. The index is held on the winters 5 screen 640, shown in Fig. 27. This data is read into the model in the Replenishment Module. I For the Generic Method,, the initial value for the forecast obtained using the generic method is simply obtained by calculating the total demand in all stores over the last week. Level = TotalSalesAllStores(Last Week) TOTAL FORECASTING 10 The total forecasting method is defined on the Forecasting screen 470, shown in Fig. 20, which is accessed from screen 330 via button 340. The total forecasting method options include: Simple Moving Average; Simple Exponential Smoothing; Holt Exponential Smoothing; Winters (Additive) Exponential Smoothing; Winters (Multiplicative) Exponential Smoothing; and Generic. 15 The parameters, which control the total forecasting, include: Simple average parameter (weeks); Exponential Smoothing Level Time Constant (range 0 to 1); Exponential Smoothing Trend Time Constant (range 0 to 1); Exponential Smoothing Trend Damping Constant; Exponential Smoothing Seasonal Time Constant (range 0 to 1); Data points to initialize with (weeks); Data points in Error Calculation (weeks); and Total demand forecast check box. 20 The parameters that may be chosen to have a default value include: Simple average parameter (weeks); Exponential Smoothing Level Time Constant (range 0 to 1); Exponential Smoothing Trend Time Constant (range 0 to 1); Exponential Smoothing Trend Damping Constant; and Exponential Smoothing Seasonal Time Constant (range 0 to 1).

WO 03/054756 PCT/IB02/05580 35 All parameters, except the total demand forecast check-box are defined via the Set Forecasting Parameters screen 440 and the default values for the parameters are assigned by clicking the default button 442 on the Set Forecasting Parameters screen 440. The total demand forecast check-box is controlled via the Define Replenishment Characteristics. 5 DEFINITIONS AND IMPACT OF PARAMETERS ON THE MODEL The following are the definitions of the parameters and their impact on the model. For the Simple Moving Average Forecasting Method, the forecast is calculated as follows. Days in the forecast = 7 x total simple parameter. The total sales for the days in the forecast are calculated as the demand minus the shortfall. The dynamic method total forecast for 10 each week in the forecast horizon is calculated as the total sales divided by the total simple parameter. For example: Total simple parameter = 5 weeks; Days in forecast = 7*5 = 35 days; 15 Demand for days in forecast = 30 units; Shortfall = 5 units; Total Sales for days in the forecast = 25 units; Forecast horizon = 3 weeks; Current week = 5b week; 20 Forecast week = 6h week; Dynamic method total forecast (5, 6) = 25/5 = 5units; Dynamic method total forecast (5, 7) = 25/5 = 5units; and Dynamic method total forecast (5, 8) = 25/5 = 5units. SIMPLE EXPONENTIAL SMOOTHING 25 For simple exponential smoothing, the current week's sales are calculated as the demand minus the shortfall. The Dynamic method total forecast for each week in the forecast horizon is calculated as (the current week sales multiplied by the total level parameter) plus (one minus WO 03/054756 PCT/IB02/05580 36 total level parameter) multiplied by the forecast for last week made for the each week in the forecast horizon past the current week. An example for the actual forecast is as follows: Total parameter level = 0.4; 5 (1- Total parameter level) = 0.6; Current weeks sales = 10 units; Forecast horizon = 3 weeks; Last week = 4 t h week; Current week = 5 th week; 10 Forecast week = 60 week; Dynamic method total forecast (4, 6) = 5 units; Dynamic method total forecast (4, 7) = 6 units; Dynamic method total forecast (4, 8) = 7 units; Dynamic method total forecast (5, 6) = 10 * 0.4 +0.6 * 5 = 7; 15 Dynamic method total forecast (5, 7) = 10 * 0.4 +0.6 * 6 = 7.6; and Dynamic method total forecast (5, 8) = 10 * 0.4 + 0.6 * 7 = 8.2. HOLT EXPONENTIAL SMOOTHING For Holt Exponential Smoothing, the sales for the current week are calculated as the demand minus the shortfall for each day in the week. 20 The Exponential Smoothing Level for the current week is calculated as total level parameter multiplied as the sales for the current week plus (one minus the total level parameter) multiplied (Exponential Smoothing Level for the week before the current week plus Exponential Smoothing Trend for the week before the current week) The Exponential Smoothing Trend for the current week is calculated as the total trend 25 parameter multiplied by the (Exponential Smoothing Level for the current week minus Exponential Smoothing Level for the week before the current week) plus (one minus the total trend parameter) multiplied by the Exponential Smoothing Trend for the week before the current week.

WO 03/054756 PCT/IB02/05580 37 A damping parameter for each week is calculated as the total damping parameter to the power of the number of weeks into the forecast horizon the current forecast week is. The Dynamic method total forecast made in the current week for each week in the forecast horizon is calculated as Exponential Smoothing Level for the current week plus (the 5 sum of the each week damping parameter multiplied by the Exponential Smoothing Trend for the current week). An example for the actual forecast is as follows: Total level parameter = 0.4; (1-Total level parameter) = 0.6; 10 Total trend parameter = 0.2; (1 -Total trend parameter) = 0.8; Total damping parameter = 2; Current weeks sales = 10 units; Exponential Smoothing Level for the 4" week = 6; 15 Exponential Trend Level for the 4 th week = 3; Exponential Smoothing Level for the 5 t h week = 0.4 * 10 + 0.6*(6 + 3) = 9.4; Exponential Trend Level for the 5th week = 0.2 * (9.4 -6) + 0.8 * 3 = 3.08; Forecast horizon = 3 weeks; Last week = 4 t h week; 20 Current week = 5 a week; Forecast week = 6 week; Dynamic method total forecast (5, 6) = 9.4 + (2 * 3.08) = 15.56; Dynamic method total forecast (5, 7) = 9.4 + (2+(2*2))*3.08 = 27.88; and Dynamic method total forecast (5, 8) = 9.4 + (2+(2*2)+(2*2*2))*3.08 = 52.52 25 WINTERS EXPONENTIAL SMOOTHING (ADDITIVE) For Winters Exponential Smoothing (Additive), the sales for the current week are calculated as the demand minus the shortfall for each day in the week. The Exponential Smoothing Level for the current week is calculated total level parameter (the sales for the current week minus the original seasonal index for the current week plus (one minus the total WO 03/054756 PCT/IB02/05580 38 level parameter) multiplied by (the Exponential Smoothing Level for the previous week plus Exponential Trend Level for the previous week). Exponential Smoothing Trend for the current week is calculated as the total trend parameter multiplied by (Exponential Smoothing Level for the current week minus Exponential 5 Smoothing Level for the previous week plus (one minus total trend parameter) multiplied by Exponential Smoothing Trend for the previous week. Exponential Smoothing Seasonal for the current week is calculated as total seasonal parameter multiplied by (the'sales for the current week minus Exponential Smoothing Level for the current week) plus (one minus total seasonal parameter) multiplied by Original Seasonal 10 Index for the current week. Each week's damping parameter is calculated as the total damping parameter to the power of the number of weeks the current forecast week is in the forecast horizon. The Dynamic Method Total Forecast made in the current week for each week in the forecast horizon is calculated as Exponential Smoothing Level for the current week plus the sum 15 of the each week damping parameter multiplied (Exponential Trend Level for the current week plus Original Seasonal Index for the week in the forecast horizon). An example for the actual forecast is as follows: Total level parameter = 0.4; (1- Total level parameter) = 0.6; 20 Total trend parameter = 0.2; (1 -Total trend parameter)= 0.8; Total seasonal parameter = 0.3; (1 -Total seasonal parameter) = 0.7; Current weeks sales = 10 units; 25 Original seasonal index for the 5 Week = 2; Original seasonal index for the 6 h Week = 8; Original seasonal index for the 7 th Week = 9; Original seasonal index for the 8 t h Week = 10; WO 03/054756 PCT/IBO2/05580 39 Exponential Smoothing Level for the 4 t h week = 6; Exponential Trend Level for the 4 h week = 3; Exponential Smoothing Level for the 5 t week = 0.4 * (10 - 2) + 0.6*(6 + 3) = 8.6; Exponential Trend Level for the 5 h week = 0.2 * (8.6 - 6) + 0.8 * 3 = 2.92; 5 Exponential Seasonal Level for the 5 t h week = 0.3*(10-8.6) + 0.7*2 = 1.82; Forecast horizon = 3 weeks; Last week = 4 t h week; Current week = 5h week; Forecast week = 6 t h week; 10 Total damping parameter = 2; Dynamic method total forecast (5, 6) = 8.6 + (2 *(2.92+8)) = 30.44; Dynamic method total forecast (5, 7) = 8.6 + (2+(2*2))*(2.92+9) = 80.12; and Dynamic method total forecast (5, 8) = 8.6 + (2+(2*2)+(2*2*2))*(2.92+10)= 189.48. WINTERS EXPONENTIAL SMOOTHING (MULTIPLICATIVE) 15 For the Winters Exponential Smoothing (Multiplicative), the sales for the current week are calculated as the demand minus the shortfall for each day in the week. The Exponential Smoothing Level for the current week is calculated as either: 1) The total level parameter multiplied to (the sales for the current week divided by the original seasonal index for the current week) plus (one minus the total level parameter) 20 multiplied by (the Exponential Smoothing Level for the previous week plus Exponential Trend Level for the previous week); or 2) Zero if the original seasonal index for the current week is zero. The Exponential Smoothing Trend for the current week as calculated as the total trend parameter multiplied by (Exponential Smoothing Level for the current week minus 25 Exponential Smoothing Level for the previous week) plus (one minus total trend parameter) multiplied by the Exponential Smoothing Trend for the previous week. The Exponential Smoothing Seasonal for the current week calculated as either: WO 03/054756 PCT/IB02/05580 40 1) The total seasonal parameter multiplied by (the sales for the current week divided by Exponential Smoothing Level for the current week) plus (one minus total seasonal parameter) multiplied by Original Seasonal Index for the current week; or 2) Zero if the original seasonal index for the current week is equal to zero. 5 Each week's damping parameter is calculated as the total damping parameter to the power of the number of weeks the current forecast week is in the forecast horizon. The dynamic method total forecast made in the current week for each week in the forecast horizon is calculated as (Exponential Smoothing Level for the current week plus the sum of the each week damping parameter multiplied by the Exponential Trend Level for the 10 current week) multiplied by the Original Seasonal Index for the week in the forecast horizon. Here the forecast needs to be equal to zero if the forecast produced is negative. An example for the actual forecast is as follows: Total level parameter = 0.4; (1- Total leVel parameter) = 0.6; 15 Total trend parameter = 0.2; (1 -Total trend parameter) = 0.8; Total seasonal parameter = 0.3; (1 -Total seasonal parameter) = 0.7; Current weeks sales = 10 units; 20 Original seasonal index for the 5" Week = 2; Original seasonal index for the 6t Week = 8; Original seasonal index for the 7t Week = 9; Original seasonal index for the 8 Week = 10; Exponential Smoothing Level for the 4 h week = 6; 25 Exponential Trend Level for the 4 t h week = 3; Exponential Smoothing Level for the 5 h week = 0.4 * (10 / 2) + 0.6*(6 + 3) = 7.4; Exponential Trend Level for the 5 t h week = 0.2 * (7.4 - 6) + 0.8 * 3 = 2.68; Exponential Seasonal Level for the 5" week = 0.3*(10/7.4) + 0.7*2 = 1.81; Forecast horizon = 3 weeks; 30 Last week = 4h week; WO 03/054756 PCT/IB02/05580 41 Current week = 5 h week; Forecast week = 6 h week; Total damping parameter = 2; Dynamic method total forecast (5, 6) = (7.4 + (2 * 2.68))*8= 102.08; 5 Dynamic method total forecast (5, 7) = (7.4 + (2+(2*2))*2.68) * 9 = 211.32; and Dynamic method total forecast (5, 8) = (7.4 + (2+(2*2)+(2*2*2))*2.68)*10= 449.2. GENERIC FORECASTING For Generic Forecasting, the average sales for the current week are calculated as the sum demand for each day in the week divided by seven. The dynamic method total forecast made in 10 the current week for each week in the forecast horizon is calculated as the average sales for the current week multiplied by the distribution center lead-time. An example for the actual forecast is as follows: Distribution center lead-time = 21 days; Weeks sales = 28 units; 15 Average weeks sales = 28/7 = 4 units; Forecast horizon = 2 weeks; Current week = 5t week; Forecast week = 6 th week; Dynamic method total forecast (5, 6) = 4*21 = 84; and 20 Dynamic method total forecast (5, 7) = 4 * 21 = 84. IMPLEMENTATION IN THE MODEL For implementation in the model, the following parameters are entered via the Forecasting Parameters screen 440, shown in Fig. 17 The values entered, including Total simple parameter, Exponential Smoothing Level 25 parameter, Exponential Trend Level parameter, Exponential Damping Level parameter, and Exponential Seasonal Level parameter, are placed on the Supply Characteristics screen 700, including the supplier to distribution center lead-time and the forecasting method.

WO 03/054756 PCT/IB02/05580 42 The default values are read from the following: Total simple parameter; Exponential Smoothing Level parameter; Exponential Trend Level parameter; Exponential Damping Level parameter; and Exponential Seasonal Level parameter. The values are read into the model by the Replenishment module. 5 STORE FORECASTING METHOD The store forecasting method may be defined on the Forecasting screen 440. The total forecasting method options are: Single Moving Average; Simple Exponential Smoothing; Holt Exponential Smoothing; Generic; External; and Hierarchical. The parameters that control the total forecasting are: Simple average parameter (weeks); 10 Store forecast parameter (weeks); Exponential Smoothing Level Time Constant (range 0 to 1); Exponential Smoothing Trend Time Constant (range 0 to 1); Exponential Smoothing Trend Damping Constant; Exponential Smoothing Seasonal Time Constant (range 0 to 1); Data points to initialize with (weeks); Data points in Error Calculation (weeks); and Total demand forecast check box. 15 The parameters that may be chosen to have a default value are: Simple average parameter (weeks); Store forecast parameter (weeks); Exponential Smoothing Level Time Constant (range 0 to 1); Exponential Smoothing Trend Time Constant (range 0 to 1); Exponential Smoothing Trend Damping Constant; and Exponential Smoothing Seasonal Time Constant (range 0 to 1). 20 All parameters, except the total demand forecast check box are defined via the Set Forecasting Parameters screen 440 and the default values for the parameters are assigned by clicking the default button 442 on the Set Forecasting Parameters screen 440. The store demand forecast check box is defined via the Define Replenishment Characteristics. In the Simple Moving Average Forecasting Method, Days in the forecast are set equal to 25 seven multiplied by the total simple parameter. The Total Sales for the days in the forecast is calculated as the demand minus the shortfall. The dynamic method store forecast for each week in the forecast horizon is calculated as the total sales divided by the total simple parameter.

WO 03/054756 PCT/IB02/05580 43 An example is as follows: Total simple parameter = 5 weeks; Days in forecast = 7*5 = 35 days; Demand for days in forecast = 30 units; 5 Shortfall = 5 units; Total Sales for days in the forecast = 25 units; Forecast horizon = 3 weeks; Current week = 5 th week; Forecast week = 6 week; 10 Dynamic method store forecast (5, 6) = 25/5 = 5units; Dynamic method store forecast (5, 7) = 25/5 = 5units; and Dynamic method store forecast (5, 8) = 25/5 = 5units. SIMPLE EXPONENTIAL SMOOTHING For Simple Exponential Smoothing, the current week sales are calculated as the demand 15 minus the shortfall. The dynamic method store forecast for each week in the forecast horizon is calculated as (the current week sales multiplied by the total level parameter) plus (one minus total level parameter) multiplied by the forecast for last week made for each week in the forecast horizon past the current week. An example for the actual forecast is as follows: 20 Store parameter level= 0.4; (1- store parameter level) = 0.6; Current weeks sales = 10 units; Forecast horizon = 3 weeks; Last week = 4 week; 25 Current week = 5 h week; Forecast week = 6 h week; Dynamic method store forecast (4, 6) = 5 units; Dynamic method store forecast (4, 7) = 6 units; Dynamic method store forecast (4, 8) = 7 units; 30 Dynamic method store forecast (5, 6) = 10 * 0.4 +0.6 * 5 = 7; WO 03/054756 PCT/IB02/05580 44 Dynamic method store forecast (5, 7) = 10 * 0.4 +0.6 * 6 = 7.6; and Dynamic method store forecast (5, 8) = 10 * 0.4 + 0.6 * 7 = 8.2. HOLT EXPONENTIAL SMOOTHING The sales for the current week are calculated as the demand minus the shortfall for each 5 day in the week. The Exponential Smoothing Level for the current week is calculated as total level parameter multiplied as the sales for the current week plus (one minus the total level parameter) multiplied (Exponential Smoothing Level for the week before the current week plus Exponential Smoothing Trend for the week before the current week). The Exponential Smoothing Trend for the current week is calculated as the total trend parameter multiplied by the 10 (Exponential Smoothing Level for the current week minus Exponential Smoothing Level for the week before the current week) plus (one minus the total trend parameter) multiplied by the Exponential Smoothing Trend for the week before the current week. The damping parameter for each week is calculated as the total damping parameter to the power of the number of weeks into the forecast horizon the current forecast week is. The 15 dynamic method store forecast made in the current week for each week in the forecast horizon is calculated as Exponential Smoothing Level for the current week plus (the sum of the each week damping parameter multiplied by the Exponential Smoothing Trend for the current week). An example for the actual forecast is as follows: Store level parameter = 0.4; 20 (1 - store level parameter) = 0.6; Store trend parameter = 0.2; (1 - store trend parameter) = 0.8; Store damping parameter = 2; Current weeks sales = 10 units; 25 Exponential Smoothing Level for the 4 week = 6; Exponential Trend Level for the 4 h week = 3; Exponential Smoothing Level for the 5 th week = 0.4 *10 + 0.6*(6 + 3) = 9.4; Exponential Trend Level for the 5 th week = 0.2 * (9.4 -6) + 0.8 * 3 = 3.08; Forecast horizon = 3 weeks; 30 Last week = 4 th week; WO 03/054756 PCT/IBO2/05580 45 Current week = 5 h week; Forecast week = 6 th week; Dynamic method store forecast (5, 6) = 9.4 + (2 * 3.08) = 15.56; Dynamic method store forecast (5, 7) = 9.4 + (2+(2*2))*3.08 = 27.88; and 5 Dynamic method store forecast (5, 8) = 9.4 + (2+(2*2)+(2*2*2))*3.08 = 52.52. GENERIC FORECASTING For Generic Forecasting, a store forecast constant (1) is calculated as the exponential of (-1/Store forecast parameter); and a store forecast constant (2) is calculated as (one minus store forecast constant (1)), i.e. the reciprocal of the first store forecasting constant. A distribution 10 center forecast constant (1) is calculated as the exponential of (-1/distribution center forecast parameter), and a distribution center forecast constant (2) is calculated as (one minus distribution center forecast constant (1)), i.e. the reciprocal of the first distribution center forecast constant. The Supply History for a store on each day in the previous week is calculated as either: 1) (The supply history for the day of the week multiplied by the store forecast constant 15 (1)) plus (the supply for the store on the day of the week multiplied by the store forecast constant (2)). If the stock for the day in the week being considered or the supply for the store on that day are greater than zero; or 2) The supply history for the day multiplied by 1.1 The Supply Week for a store is calculated sum of each day in the current week as either: 20 1) The supply for the store that day if the stock for the day in the week being considered or the supply for the store on that day are greater than zero; or 2) The supply history for the store that day. The Supply history week is calculated as either: WO 03/054756 PCT/IB02/05580 46 1) If the supply week is greater than zero, the supply history week multiplied by the store forecast constant (1) plus the supply week multiplied by the distribution center forecast constant (2). 2) Otherwise, the supply history week multiplied by 1.1. 5 The dynamic method store forecast is calculated as the supply history week for the store for the week being considered, for each week in the forecast horizon. An example for the actual forecast is as follows: For store 1: Store forecast parameter = 5; 10 Store forecast constant (1) = e (-/5); Store forecast constant (2) = (1 - e('); Distribution center forecast parameter = 6; Distribution center forecast constant (1) = e (-/6) Distribution center forecast constant (2) = (1 - e 1/6)); 15 Current week = 3 rd week; Initialized Supply history Supply History (1,1) = 1; Supply History (1,2) = 2; Supply History (1,3) = 3; 20 Supply History (1,4) = 4; Supply History (1,5) = 5; Supply History (1,6) = 6; Supply History (1,7) = 7; and Supply History Week = 10. 25 For day 15: Supply (1,15) = 2 units; Stock (1,15) = 6 units; and Supply History (1,1) = 1* e ( 1 5) +2*(1 -e(-) WO 03/054756 PCT/IB02/05580 47 For day 16: Supply (1,16) = 3 units; Stock (1,16) = 5 units; and Supply History (1,2) = 2* e ' 1/5) +3*(1 - e(/.

5)) 5 For day 17: Supply (1,17)= 0 units; Stock (1,17) = 0 units; and Supply History (1,3) = 3*1.1. For day 18: 10 Supply (1,18) = 1 units; Stock (1,18) = 0 units; and Supply History (1,3)= 4* e(- /s) +1 *(1 - e(-5)). For day 19: Supply (1,19) = 4 units; 15 Stock (1,19) = 3 units; and Supply History (1,4)= 5* e (- /s) +4*(1 - e(u5)). For day 20: Supply (1,20) = 5 units; Stock (1,20) = 1 units; and 20 Supply History (1,5) = 6* e(- /s) +5*(1 - e() For day 21: Supply (1,21) = 2 units; Stock (1,21) = 2 units; Supply History (1,6) = 7* e (-1/5) +2*(1 - e (-1/5)); 25 Supply Week =2+3+(3*1.1)+1+4+5+2 = 20.3; and Supply history week = 10 * e (u s) +20.3*(1 - e(/6).

WO 03/054756 PCT/IB02/05580 48 The Dynamic method store forecast for each week in the forecast horizon = Supply history week. HIERARCHICAL FORECASTING For Hierarchical Forecasting, the Total store forecast for a week in the forecast horizon is 5 calculated as the sum for all stores forecasts made in the current week for the week in the forecast horizon. The new store forecast for the forecast week in the forecast horizon is calculated as either: 1) Zero, if the total store forecast is zero; or 10 2) The total forecast made in the current week for the forecast week multiplied by the ratio of the original store forecast for the forecast week and the total store forecast for the forecast week. As an example, the Forecast horizon = 3 weeks, the Current week = 2 nd week, and the number stores in the chain = 2. 15 The Store forecast method with simple moving average generates: Store 1: Forecast for week 3 = 3 units Forecast for week 4 = 8 units Forecast for week 5 = 12 units 20 Store 2: Forecast for week 3 = 4 units Forecast for week 4 = 9 units Forecast for week 5 = 6 units WO 03/054756 PCT/IB02/05580 49 The total forecast method with Winters additive generated the following chain forecasts: Forecast for week 3 = 14 units; Forecast for week 4 = 20 units; and Forecast for week 5 = 19 units. 5 Total store forecast for week 3 = 3+4 = 7; Total store forecast for week 4 = 8+9 = 17; Total store forecast for week 5 = 12+6 = 18; The Hierarchical implies that the store forecasts are modified such that: For store 1: 10 New forecast for week 3 = 14 *(3/7) =6 units; New forecast for week 4 = 20 *(8/17) =9.4 units; and New forecast for week 5 = 19*(12/18) =12.7 units. For store 2: New forecast for week 3 = 14 *(4/7) =8 units; 15 New forecast for week 4 = 20 *(9/17) =10.6 units; and New forecast for week 5 = 19*(6/18) = 6.3 units. IMPLEMENTATION IN THE MODEL For implementation in the model, all of the parameters are entered via the Forecasting Parameters screen. The values entered are placed on the Supply Characteristics screen 700 as 20 follows: Total simple parameter; Generic simple parameter; Exponential Smoothing Level parameter; Exponential Trend Level parameter; Exponential Damping Level parameter; and Exponential Seasonal Level parameter. The default values are read as follows: Total simple parameter; Exponential Smoothing Level parameter; Exponential Trend Level parameter; Exponential Damping Level parameter; 25 and Exponential Seasonal Level parameter.

WO 03/054756 PCT/IB02/05580 50 DISTRIBUTION CENTER REPLENISHMENT The distribution center replenishment method is defined on the Replenishment Methods screens 420 and 421, which are accessed via the Replenishment Characteristics screen 330 via button 336. 5 The distribution center replenishment method options are: Min-Max; Adelsten Specific; and Dynamic method. The Adelsten Specific and Dynamic Methods may only be employed for the distribution center when the same method is used for store replenishment. The Min Max method may be used at the distribution center for all store replenishment methods. 10 The parameters that control the distribution center replenishment are: Distribution center Safety Stock (days cover); Distribution center review time (days); Supplier to distribution center lead-time (days); Supplier minimum order quantity (units); and Total demand forecast check box. All parameters are defined via the screens 420 and 421, in Fig. 14a and 14b. 15 According to the Min-Max distribution center replenishment method, the precise operation of the Min-Max distribution center method varies according to whether the Total Demand Forecast check box is selected. The minimum distribution center stock is calculated as: a) When the distribution center forecasting is disabled: [Supplier lead time (days) + the number of days of safety stock required] x the average 20 daily demand. Minimum = [SupplierLead_Time + Safety_Stock] x Average_Daily Demand, b) When the distribution center forecasting is enabled: Number of days of safety stock x the average daily demand.

WO 03/054756 PCT/IB02/05580 51 Minimum = Safety_Stock x Average Daily_Demand The maximum distribution center stock is calculated as: The minimum distribution center stock (calculated above) + half of the supplier minimum order quantity. 5 Maximum = Minimum + Supplier MinimumOrder Quantity / 2 If the current distribution center stock level, including any stock in the supply chain between Supplier and distribution center, is less than the minimum allowable stock then an order needs to be generated. That is, the distribution center orders more stock from suppliers. If Total Demand Forecasting is enabled, then the sales forecast for the days between the 10 current day and the day that is equal to (DCReview_Time + SupplierDCLead_Time) days in the future is also used in determining whether an order needs to be raised. The determination of whether to order stock depends on whether DistributionCenter Stock plus Stockin Transit is less than the Minimum (optionally, plus the forecast). If so, then more stock is ordered. The quantity ordered should be enough to raise the distribution center stock to the 15 maximum level. Thus, the order quantity is calculated as: Order Quantity = Maximum - Distribution_Center_Stock - Stock in Transit [+ Forecast] This order quantity is then rounded up to the nearest multiple of the store minimum order quantity. This value may be restricted if limits are placed on supplier availability. 20 For example: Supplier Minimum Order Quantity = 100 units; Supplier's Lead Time = 14 days; Distribution Center stock = 200 units; Safety stock = 28 days; 25 Stock in chain = 50 units; and WO 03/054756 PCT/IB02/05580 52 Average daily demand = 10 units. With the forecast enabled: Forecast = 100 units Minimum = 28 x 10 = 280 units 5 Maximum = 280 + 100/2 = 330 units Minimum + Forecast = 380 units > DistributionCenterStock + StockinChain = 200 units. Thus, the system generates an order because the distribution center stock plus the stock in chain is less than the minimum amount plus the forecast amount. The order amount is 10 determined as follows: Order Quantity = Maximum - DistributionCenterStock - Stock In Chain + Forecast. = 330 - 200 -0 +100 = 230 The order quantity may be rounded up by a predetermined amount, e.g., to the next even 15 hundred (or any other number) units, to be 300 units. With the forecast disabled: Minimum = (28 + 14) x 10 = 420 units Maximum = 420 + 100/2 = 470 units Because the Minimum (420 units) is greater than the Distribution_Center_Stock + 20 Stock in Chain (200 units), an order is generate as follows: Order Quantity = Maximum - DistributionCenter_Stock - StockIn Chain = 470 - 200 - 0 = 270 After rounding the order quantity up, the order quantity is 300 units.

WO 03/054756 PCT/IB02/05580 53 ADELSTEN SPECIFIC METHOD The Adelsten Distribution Center method may be used when the Adelsten method has been selected for the store. The distribution center is replenished by looking at each of the stores individually, and using the Adelsten specific method for each store. However, the distribution 5 center lead-time and review time parameters are used to calculate individual store order quantities. It then sums the positive order quantities to obtain the distribution center order quantity. An Adelsten constant' is defined in the Replenishment Methods screen and is written to the Supply Characteristics screen 700. The Minimum Stock for each store group are defined on O10 the Replenishment Methods screen 420 and 421. The Average Sales (actual demand - shortfall) over a period of weeks is calculated. The Variance is calculated by finding the difference between the maximum weekly sales during the current week sales and the average sales for the store. The Total_Time is the (Distribution_CenterReviewTime + 15 Distribution_Center_Lead_Time) in days. The stock level for the store is then calculated as: Target Stock = Minimum Stock + (Average Sales + Variance x Adelsten_Constant) x (Total Time 7 The amount required by the store is the Target Stock minus the current store stock minus the stock in the supply chain. 20 Store Need = Target Stock - Store Stock -Store Supply Chain This store need is then rounded up to the nearest store order multiple.

WO 03/054756 PCT/IB02/05580 54 The amount required by the distribution center is then the sum of the positive amounts required by each store minus the current distribution center stock and the distribution center chain. DC Need = Positive Store Needs - DC _ Stock - DC Supply Chain All Stores 5 This value is then rounded up to the nearest Supplier Minimum Order Quantity Multiple. The value is then restricted if limits are placed on the supplier availability. Example: Variable Value0 Sales this week 10 No. of weeks under consideration 5 Maximum weekly sales over last 5 16 weeks Sales for past 5 weeks 40 Average sales 8 Variance 8 Minimum Stock 2 Adelsten Constant 0.7 Supplier to DC Lead Time 7 days DC Review Time 14 days WO 03/054756 PCT/IB02/05580 55 Supplier Minimum Order Quantity 100 Store Minimum Order Quantity 1 Target Stock = MinimumStock + (Average Sales + Variancex Adelsten_Constant) x (TotalTim e 7 =2 +(8+7x0.7 ) ( 1 7 = 40.7, Store Need = Target Stock - Store Stock-Store Supply Chain =40.7- 20 - 10 = 10.7 Store Need = 11 This is repeated for each store, and the positive store needs summed to get a total need. 5 Sum of Store Needs = 439 Distribution Center Stores Required = 439 units Present Distribution Center stock = 300 units Present units in the chain = 10 units 10 Distribution Center requirement = 439 - 300 - 100 = 39 units Distribution Center Requirement rounded up to nearest Supplier Minimum Order Quantity multiple = 100 units. DYNAMIC METHOD For the Dynamic Method, shipments are calculated as the total amount of units in the 15 store chain for the store over the store lead-time period. Forecast stock is calculated as the current stock plus the shipments minus the forecasted sales for the next lead time period.

WO 03/054756 PCT/IB02/05580 56 An index W is calculated by (1 - service level for the store group) multiplied by the forecast sales for the period of the review time after the store lead time divided by the store forecast error. The difference in W is calculated as the difference in value of the index W and the index greater than W in the lookup table. 5 The difference in Z is calculated as the difference in the Z values that correspond to the index W and the one with an index greater than W. A new Z value is calculated as the Z value corresponding to W + (W - the closest W indexed in the lookup table) x (Difference in Z) / (Difference in W). Safety Stock is calculated as the new Z value multiplied by the store error. 10 The Forecast week is calculated as the current week + (store lead time + the inventory selling days) / 7 to convert to weeks. Dynamic Lost Sales is calculated as (1-dynamic lost sales factor) x (forecast - stock). The Order Point is calculated as either (dependent on the calculated safety stock being larger or smaller than the presentation stock given in the 'replenishment parameters screen 420): 15 1) Safety stock + the forecast sales for the lead time + forecast sales for the review time, or 2) Presentation stock + the forecast sales for the lead-time + forecast sales for the review time. The Order upto point is calculated as either (dependent on the calculated safety stock 20 being larger or smaller than the presentation stock given in the 'replenishment parameters screen 420): 1) Safety stock + the forecast sales for the lead time + forecast sales for the inventory selling days, or WO 03/054756 PCT/IB02/05580 57 2) Presentation stock + the forecast sales for the lead-time + forecast sales for the inventory selling days. A Reorder quantity is calculated as the order upto point - the forecast stock - the Dynamic Lost Sales. A required amount for the distribution center is calculated as the sum of the 5 stores reorder quantities - the distribution center stock and the distribution center chain. The following is an example: Shipments for the chain 30 units Forecast sales for current week to the end of 20 units the lead time Forecast sales for current week to the end of 30 units the (lead time plus review time) Forecast sales for current week to the end of 35 units the (lead time plus inventory selling days) Forecast sales for the end of the lead time to 15 units the end of the inventory selling days Forecast sales for the end of the lead-time to 10 units the end of the review time Current stock at the store 5 units Forecast stock 5 + 30 - 20 = 15 units Store Error 2 Dynamic Lost Sales factor 0.25 Service level 0.5 Presentation stock 3 Index W = (1 - 0.5)* (10/2) WO 03/054756 PCT/IB02/05580 58 =2.5 Lookup W (72) 2.5 Lookup W (73) 2.8 Lookup Z (72) 0.002 Lookup Z (73) 0.001 Difference inW = 2.8 - 2.5 = 0.3 Difference in Z = 0.001 -0002 = - 0.001 New index Z = 0.002 +(2.5 2.5)*(-0.001/0.3) = 0.002 Safety Stock = 0.002*2 = 0.004 New Safety stock = 1 Forecast sales for current week to the end of the lead-time are greater than forecast stock. Thus, dynamic lost sales = (1 - 0.25) x (20-15) = 3.75 Because Presentation Stock is greater than Safety Stock, Presentation Stock is used in 5 calculations, such as: Order point = 3 +20+10 = 33; and Order upto point = 3+20+15 = 38 Recommended Order Quantity for Store = 38 -15-3.75 = 19.25 units WO 03/054756 PCT/IB02/05580 59 Round Recommended Order Quantity up to nearest store minimum order quantity multiple: Recommended Order Quantity for Store = 20 units. IMPLEMENTATION IN THE MODEL 5 For implementation in the system, all of the parameters are entered on the Supply Characteristics screen 700 as follows: Distribution center review; Supplier to distribution center lead-time; Supplier minimum order Quantity; Total demand forecast check box. The values are linked to other cells on the Excel® sheets: A calculation of the distribution center minimum order quantity (Weeks supply) on 10 Supply Characteristics screen 700; a calculation of distribution center minimum order quantity (Units); and a calculation of distribution center initial stock. STORE REPLENISHMENT For store replenishment the min - max parameters are defined in the Replenishment Methods screen 420. 15 The Store replenishment methods include: Min-Max; Constant; Mercatus Ideal; Adelsten Specific; and Dynamic Method. The parameters controlling store replenishment depend on the chosen method: Store min (weeks supply or actual supply) for each individual store group; Store max (weeks supply, actual supply or multiple of min) for each individual store 20 group; or Weeks cover for each individual store group. All parameters are defined via the Replenishment Methods screen 420 and 421. In one embodiment, the Min-Max replenishment method uses the following process. If the stock level in a store falls below a set minimum quantity, then extra stock is ordered to raise 25 the stock to the maximum value. The minimum stock level may be set in two ways. First, WO 03/054756 PCT/IB02/05580 60 minimum stock level may be a number of days cover, which is translated into units using the store forecast. Second, minimum stock level may be a fixed number of units. Each store group may have a different value for the parameter. The maximum stock level may be calculated in three ways. First, the maximum stock level may be a set number of 5 multiples of the minimum stock level. Second, the maximum stock level may be a number of days cover, which is translated into units using the store forecast. Third, the maximum stock level may be a fixed number of units. Each store group may have a different parameter. The following is an example: Parameter Value Store minimum 5 units stock Store Maximum = 3 times minimum Stock stock 3 x5 = 15 units Stock in Store 2 Stock in the 2 Store Chain 10 Stock on Hand= Stock in Store + Stock in Chain =2 +2 = 4 units This is less than the minimum stock level, so an order is generated.

WO 03/054756 PCT/IB02/05580 61 Order Quantity = Maximum Stock - Available Stock = 15-4 = 11 units. CONSTANT STORE REPLENISHMENT METHOD 5 In one embodiment, the Constant replenishment method follows the following process. If the stock level falls below a constant minimum level, then stock is ordered to bring stock level back to the constant level. The constant minimum stock level may be defined in two ways. First, the constant minimum stock may be a number of days cover, which is translated into units using the store forecast. Second, the constant minimum stock may be a fixed number of units. 10 The following is an example: Parameter Value Store minimum 7 units stock Stock in Store 2 Stock in the Store 2 Chain Stock on Hand= Stock in Store + Stock in Chain =2 +2 = 4 units This is less than the minimum stock level, so an order is placed. 15 Order Quantity = Constant Stock - Available Stock =7-4 = 3 units WO 03/054756 PCT/IB02/05580 62 One further criterion applies to the constant method. If the Store Minimum Order Quantity is 2 units or more, then order quantities need to be in multiples of this minimum order quantity. Hence, in the above example the order quantity would be rounded up to 4 units. MERCATUS IDEAL STORE REPLENISHMENT METHOD 5 The Mercatus store method is similar to the Min Max method, with one extra condition. The minimum stock level may be set in two ways. First, the minimum stock level may be a number of days cover, which is translated into units using the store forecast. Second, the minimum stock level may be A fixed number of units. Each store group may have a different value for the parameter. 10 The maximum stock level may be calculated in three ways. First, the maximum stock level may be a set number of multiples of the minimum stock level. Second, the maximum stock level may be a number of days cover, which is translated into units using the store forecast. Third, the maximum stock level may be a fixed number of units. The Ideal Stock level is calculated as a number of days cover, which is translated into 15 units using the store forecast. The Order upto level is calculated as: 1) The maximum stock if the ideal stock level is greater than the maximum stock; 2) The minimum stock if the ideal stock level is less than the minimum stock; or 3) The ideal stock level otherwise. 20 The following is an example: Parameter Value Store minimum stock 5 units Store maximum Stock 15 units WO 03/054756 PCT/IB02/05580 63 Mercatus Days Cover 10 days Weekly Sales Forecast (1 wk 3.6 units ahead) Weekly Sales Forecast (2 wks 5.6 units ahead) Stock in Store 2 Stock in the Store Chain 2 Ideal Stock Level = 10 days cover = Forecast (1 week ahead) + Part of Forecast (2 weeks ahead) = 3.6 + (3/7) x 5.6 = 6 units 5 Thus, the order up to point is 6 units because that is the number of units needed to provide 10 days cover for sales. The order quantity is calculated using the simple formula: Order Quantity = Order up to Point - Current Stock =6-4 = 2 units. 10 ADELSTEN SPECIFIC REPLENISHMENT METHOD For the Adelsten Specific Replenishment Method, the minimum stock level may be set in two ways. First, the minimum stock level may be a number of days cover, which is translated into units using the store forecast. Second, the minimum stock level may be a fixed number of units. Each store group may have a different value for the parameter. 15 The Minimum Stock for each store group is defined on the Replenishment Methods screen 421, shown in Fig. 14a. The Average Sales (actual demand - shortfall) over a period of weeks is calculated. The Variance is calculated by finding the difference between the maximum weekly sales during the current week sales and the average sales for the store.

WO 03/054756 PCT/IB02/05580 64 The Total Time = (ReviewTime + Store Lead-Time) in days. The stock level for the store is then calculated as: Target Stock = Minimum Stock + (Average Sales + Variance x Adelsten_Constant) x Total Time 7 The amount required by the store is the Target Stock minus the current store stock minus 5 the stock in the supply chain., Store Need = Target Stock - Store Stock -Store Supply Chain This store need is then rounded up to the nearest store order multiple. An example is as follows: Variable Value Sales this week 10 No. of weeks under consideration 5 Maximum weekly sales over last 5 weeks 16 Sales for past 5 weeks 40 Average sales 8 Variance 8 Minimum Stock 2 Adelsten Constant 0.7 WO 03/054756 PCT/IB02/05580 65 Store Lead Time 7 days Review Time 14 days Supplier Minimum Order Quantity 100 Store Minimum Order Quantity 1 Target Stock = Minimum Stock + (Average Sales + Variance x AdelstenConstant) x (Total

T

im e - 7) =2 +(8+7x0.7)x2 = 40.7 Store Need = Target Stock - Store Stock -Store Supply Chain =40.7- 20 - 10 = 10.7 Store Need = 11. DYNAMIC METHOD STORE REPLENISHMENT 5 For Dynamic Store Replenishment, shipments are calculated as the total amount of units in the store chain for the store over the store lead-time period. Forecast stock is calculated as the current stock plus the shipments minus the forecasted sales for the next lead-time period. An index W is calculated by (1 - service level for the store group) multiplied by the forecast sales for the period of the review time after the store lead time divided by the store forecast error. A 10 difference in W is calculated as the difference in value of the index W and the index greater than W in the lookup table held in screens 'Lookup Table'. A difference in Z is calculated as the difference in the Z values that correspond to the index W and the one with an index greater than W. A new Z value is calculated as the Z value corresponding to W plus (W - the closest W indexed in the 'lookup table') multiplied WO 03/054756 PCT/IB02/05580 66 (Difference in Z) divided (Difference in W). Safety Stock is calculated as the new Z value multiplied store error. Forecast week is calculated as the current week plus (store lead time plus the inventory selling days), divided by seven to convert to weeks. Dynamic Lost Sales is calculated as (1 - dynamic lost sales factor) x (difference between the forecast and the stock). 5 The Order Point is calculated as either (dependent on the calculated safety stock being larger or smaller than the presentation stock given in the 'replenishment parameter' screen 420): 1) Safety stock + the forecast sales for the lead time + forecast sales for the review time; or 2) Presentation stock + the forecast sales for the lead-time + forecast sales for the review 10 time. The Order upto point is calculated as either (dependent on the calculated safety stock being larger or smaller than the presentation stock given in the 'replenishment parameters screen 420): 1) Safety stock + the forecast sales for the lead time + forecast sales for the inventory 15 selling days; or 2) Presentation stock + the forecast sales for the lead-time + forecast sales for the inventory selling days. The required amount is calculated as the order upto point minus the forecast stock minus the Dynamic Lost Sales. 20 For example: Shipments for the chain = 30 units; Forecast sales for current week to the end of the lead time = 20 units; Forecast sales for current week to the end of the (lead time + review time) = 30 units; Forecast sales for current week to the end of the (lead time + inventory selling days) = 35 25 units; WO 03/054756 PCT/IB02/05580 67 Forecast sales for the end of the lead time to the end of the inventory selling days = 15 units; Forecast sales for the end of the lead time to the end of the review time = 10 units; Current stock at the store = 5 units; 5 Forecast stock = 5 + 30-20 = 15 units; Store Error = 2; Dynamic Lost Sales factor = 0.25; Service level for store group 1 = 0.5; Promotion stock for store group 1 = 3; 10 Index W = (1 - 0.5)* (10/2) = 2.5; Lookup W (72) = 2.5; Lookup W (73) = 2.8; Lookup Z (72) = 0.002; Lookup Z (73) = 0.001; 15 Difference in W= 2.8 - 2.5 = 0.3; Difference in Z = 0.001 -0.002 = -0.001; New index Z = 0.002 +(2.5 - 2.5)*(-0.001/0.3) = 0.002; Safety Stock = 0.002*2 = 0.004; Round up to nearest integer: 20 New Safety stock = 1; As Forecast sales for current week to the end of the lead-time > forecast stock. Dynamic Lost Sales = (1 - 0.25)* (20-15) = 3.75; As Promotion Stock is greater than the safety stock; Order point = 3 +20+10 = 33; 25 Order upto point= 3+20+15 = 38; and Required quantity for the store = 38 -15-3.75 = 19.25 units. IMPLEMENTATION IN THE MODEL Most of the parameters are entered via the parameters screen 406, shown in Fig. 12. The values entered are placed on the Supply Characteristics screen 700: Number of 30 weeks; Adelsten constant; Store group minimum; Store group maximum; Mercatus weeks cover; Store lead-time; Presentation stock; Inventory selling days; Service level; and Lost Sales factor.

WO 03/054756 PCT/IB02/05580 68 A calculation of the Store group minimum for constant and Adelsten replenishment on Supply Characteristics screen 700 is performed. The values are read into the system. ALLOCATIONS: PUSH BASED STRATEGIES For all push based replenishment strategies the number of units that should be allocated 5 to each store, and when these allocations should be made may be defined. The flow options in the model, to which these allocations apply, are: Single Push, Push in Waves, and Controlled Push & Push-Pull. Using the Allocations' screen 408, the size and timing of any allocations may be determined. The Allocations' screen 408 is accessed via Set Allocations button 334. 10 A maximum of six pushes may be defined using the screen. The timing of any push is also controllable from this screen. The user is allowed to set both the week and day of each push. This date is the day on which the allocated stock is dispatched from the warehouse, or distribution center. The stock is not available in the stores until it has been delivered, i.e. the store lead-time after the dispatch day. The first time a push may be performed is on day 1 of 15 week 1, i.e. the first day of the simulation. The push currently being defined is also shown on the screen. When this is changed, the list boxes, which display the quantities pushed to each store within a store group, and the text boxes, which display the total pushed (in the current push) to each store group are updated to reflect the values for the newly selected push. 20 Once a push number and a method of allocation are selected, the size of the allocation may be set. There are two methods of determining the allocation size. Either a specific number of units may be allocated to each store, or a number of weeks' cover may be allocated to a store. The screen allows the user to choose which allocation method to implement. The weeks cover method may not be used for the first push. 25 If the allocation method chosen is Set Individual Allocation then the for each of the six store groups, the screen shows, and allows the modification of the number of units pushed to each store within the store group. It also shows the total number of units pushed to each store WO 03/054756 PCT/IB02/05580 69 group in this push. This value is calculated by multiplying the number of stores in the store group by the number allocated to each store. If the allocation method chosen is Use Weeks' Cover, then the screen allows the user to control the number of weeks cover of stock to be pushed. This number is applied to all store 5 groups. The number of units to be pushed to each store is calculated by multiplying the average weekly demand for each store group by the number of weeks' cover. This is then rounded up to the nearest integer. The average weekly demand for each store group is calculated from the beginning of the simulation, until the time of the push. This means that the maximum amount of information available at the time of the push is used when determining the push quantities. The 10 screen also shows the number of units that will be pushed to each store, and the total number of units pushed to each store group for this push. Once again, this is simply the number of stores in the store group multiplied by the number allocated to each store. The screen also contains a button that completely resets all the allocations to zero. In some cases, the number of units allocated to the stores will not exactly match the number of 15 units purchased. In this case, there may be some remaining units. The Allocations' screen 408 allows these remaining units to be allocated. There are two methods of allocating the remaining units; the method to be used is controlled by a radio button. The method under use is written to the Supply Characteristics screen 700. The first method, Allocate Remainder to Push 1, allocates all of the remaining units 20 to the first push. Starting with the largest store, one extra unit is allocated to each store until all of the remaining units are used up. If, after an extra unit has been allocated to each store, there are still some remaining items, the process is repeated and a further extra unit is allocated to the largest store, and so on until all the units are used up. The second method, Spread Remainder across Pushes, allocates the remaining units to all 25 of the pushes, if there is a sufficient quantity. Again, starting with the largest store, and moving through the stores in descending order, one extra unit is allocated to each store in the first push. If, after the extra units have been added to the first push, there is still a remaining quantity, we proceed to the next push, and starting with the largest store, and proceeding in descending order, a further extra unit is allocated in the second push. This process is repeated, for each push, until 30 all of the remaining units are used up.

WO 03/054756 PCT/IB02/05580 70 The number of units available for allocation, the Initial Purchase Quantity 331, is set on the Replenishment Characteristics screen 330, and is written to the Supply Characteristics screen 700. The screen also keeps a running total of the number of units that have been allocated, and the number of units that remain to be allocated. The screen will not allow you to allocate more 5 units than you have purchased. If this is attempted, the screen will not allow the data to be committed, until the allocation quantity is less than, or equal to, the amount available for allocation. PULL-BASED STRATEGIES Setting the size of allocations does not apply to most pull based scenarios, with the 10 exception of push pull. For push pull strategies, an initial allocation is made to each store, before switching to pull based, or replenishment. However, in some cases the replenishment methods require there to be some form of allocation. For example, if the amount required by each store based on some replenishment algorithm is calculated, and the total of these store needs is greater than the total amount of stock available in the warehouse, then the order quantity for every store 15 may not be satisfied. As a result the store order quantities need to be recalculated before the stock is distributed. Once all the store order quantities have been calculated for a particular day, the total order quantity is compared with the available stock in the distribution center. If the distribution center has sufficient stock to satisfy all of the stores orders, then each 20 store gets the amount ordered. If the distribution center does not have sufficient stock, then the stock between the stores should be rationed. The fraction of the orders that could be satisfied with the current distribution center stock level is calculated. Fraction = DC Stock Total Required 25 The individual store order quantities are then multiplied by this fraction and rounded up to the nearest store minimum order quantity multiple, to get a new order quantity. If the new WO 03/054756 PCT/IB02/05580 71 total order quantity is less than, or equal to, the distribution center stock level, then each store receives its newly calculated order quantity. If the new total order quantity is still greater than the available distribution center stock, then the order quantity needs to be reduced further before the final orders are placed. This is 5 achieved by subtracting one store minimum order quantity multiple from the order quantity of the first store (the smallest store). If the total order quantity is still greater than the distribution center stock, then a store minimum order quantity multiple is removed from the order quantity of the next store. This process is repeated until the total order quantity is equal to the available distribution center stock. Once this condition is satisfied, the final order quantities are placed 10 and the goods are dispatched to the distribution center. Example: Parameter Value Store Minimum Order Quantity 1 Available DC Stock 18 Store Order Quantities Value Store 1 5 Store 2 5 Store 3 7 Store 4 3 Store 5 9 Total 30 WO 03/054756 PCT/IB02/05580 72 Fraction = 18/30 = 0.6 Store Order Quantities Old New Value Value Store 15 =0.6x 5 =3 Store 2 5 =0.6 x 5 =3 Store3 7 =0.6x7 =4.2 =5 Store4 3 =0.6x3 = 1.8 =2 Store 5 10 =0.6x 10 =6 Total 29 19 The total order quantity (19) is greater than the total available stock (18). As a result, the order quantity for a first store is reduced. The order quantity for store number one is now: 3 - 1 =2 Store Order Quantities Final Value Store 1 2 Store 2 3 Store 3 5 Store 4 2 Store 5 6 Total 18 WO 03/054756 PCT/IB02/05580 73 These rationing rules may be found in the Replenishment Methods module, and the Store Rationing procedure. CROSS DOCKING When Cross-Docking is selected as the replenishment method, there are two methods of 5 allocating available stock. One method involves limiting the supplier availability, and possibly the amount of received by each store, the other allows stores to receive exactly the stock they require. A check box 342 entitled: Limit Supplier Availability appears on the Replenishment Characteristics screen 330, shown in Fig. 11 10 If the checkbox 342 is left unchecked, then all stores will receive exactly the quantity they require. If the checkbox 342 is checked, then the supplier limiting criteria are used to control the stock available from the supplier. These criteria may be controlled via the Supplier Availability screen 350, shown in Fig. 23, accessed via the Define Supply Availability Characteristics button 322. This check box is written to the Supply Characteristics screen 700. 15 Two methods may be used to limit the supplier availability. Method one allows only a predefined percentage of an order to be satisfied. For example if 500 units are required by the stores, but the supplier availability is restricted to 65%, then the maximum number of units available is 65% of 500 units = 325 units. The other method allows the order quantity to be restricted to at most a maximum value. 20 For example, if the stores require 500 units, as determined by the current replenishment algorithm, but the supplier availability is restricted to 300 units, then the amount available will be only 300 units. Alternatively, if there is no limit to the supplier availability, and the stores require 500 units in total, then 500 units will be available to the stores. If no restriction has been placed on the supplier availability, or the restricted value is 25 large enough to satisfy all the store requirements, then the stores receive their full requirement. If the amount available to the stores is less than the total order quantity then the available stock should be rationed amongst the stores.

WO 03/054756 PCT/IB02/05580 74 This is achieved by subtracting one store minimum order quantity multiple from the order quantity of the first store (the smallest store). If the total order quantity is still greater than the available stock, then a store minimum order quantity multiple is removed from the order quantity of the next store. This process is repeated until the total order quantity is equal to the 5 available stock. Once this condition is satisfied, the final order quantities are placed and the items dispatched to the store. Parameter Value Store Minimum Order Quantity 1 Available Stock 12 Store Order Quantities Value Store 1 1 Store 2 0 Store 3 2 Store 4 3 Store 5 1 Store 6 1 Store 7 2 Store 8 1 Store 9 1 Store 10 2 WO 03/054756 PCT/IB02/05580 75 Total 14 Again, not enough stock is available, so Store #1's order quantity is reduced, and the new store #1 order quantity = 0. Store Order Quantities Value Store 10 Store2 0 Store3 2 Store4 3 Store5 1 Store6 1 Store 7 2 Store 8 1 Store9 1 Store 10 2 Total 13 WO 03/054756 PCT/IBO2/05580 76 Again, not enough stock is available, so store #3's order quantity is reduced by one. Store Order Quantities Value Store 1 0 Store 2 0 Store 3 1 Store 4 3 Store 5 1 Store 6 1 Store 7 2 Store 8 1 Store 9 1 Store 10 2 Total 12 RESULTS The results obtained by the model may be saved. In illustrative embodiments, the model 5 has three means of outputting the results of simulations. The first method is by plotting a graph of the main variables, such as Store Inventory, distribution center Inventory and Lost Sales. The other two methods output a series of numerical results to predefined spreadsheets.

WO 03/054756 PCT/IB02/05580 77 The plot inventory and availability button 514 in the main menu 300 should be used after the modeling simulation has been run. The system will then use the results of the modeling calculation to generate graphical representation of the result of the most recent run. The modeling parameters used, the sales, demand, inventory, delivery, service level, and lost sales 5 information for each modeling run are recorded by selecting the results archive button 516. When the user selects the push/pull analysis button 518, the figures used to calculate the GPADE and PI are recorded and stored. When the run was performed, the flow path, the revenues, and the costs at each stage of the supply chain are recorded. The push/pull analysis and results archive buttons mnay be used in conjunction to determine which parameters were 10 used to achieve specific modeling results. GRAPHS The system may use three graphs to output its results, an Actual Graph, a Demand Graph, and a Supply Graph. Only two spreadsheet charts are used to display the three graphs, Supply Plot 720, shown in Fig. 21 and Demand Plot 710 shown in Fig. 15. The Actual Graph 15 and the Supply Graph plot similar information but with different data, hence both graphs use the same spreadsheet chart. The graphs are summarized in the following sections. All the graphs may have the same look and feel. Each graph may be plotted at three different levels. That is, the information may be displayed for an individual store, a collection of stores aggregated together (a store group) or all stores aggregated together. The idea of store 20 groups was explained in section 3. The user may switch between the different plot levels, i.e. individual store, store group or all stores. The user may switch between different store groups, i.e. Store Group 1 to Store Group 2. It is also possible to move back and forth between stores. The arrow buttons allow the user to plot the details for different stores in a particular store group. 25 ACTUAL GRAPH Once the external data file has been prepared, the graph of the actual data 710 may be plotted, as shown in Fig. 15. This graph 710 is accessed via the Plot Demand button 522 after selecting the base case analysis button 512 on the main menu 300. On pressing this button 522, the model asks whether an external should be used. The variables that may be plotted in this WO 03/054756 PCT/IB02/05580 78 graph are store inventory, consumer demand, lost sales, and days cover. The demand is calculated across the period. The demand is then plotted on a daily, or weekly basis depending upon the users preference, and presented along with a sum of the total and average daily demand across the period. The graph 712 provides the user with a reference point for viewing the 5 demand profile for the period under study. Alternatively, the user may select the plot current supply chain button after selecting the base case analysis button 512 on the main menu 300. As shown in Fig. 21, the graph 720 is presented to the user. The graph 720 includes the consumer demand, the lost sales, the store inventory, the regional distribution center inventory, the store days cover, the NDC inventory, 10 the days inventory is dispatched to the stores, the days inventory that is dispatched to the regional distribution center, and the days inventory that is dispatched to the NDC. In addition the graph 720 also includes non-graphical data, namely, the representative product number, the average demand chain wide every day, the average inventories by day, the deliveries to the store from the distribution center, the deliveries to the distribution center from 15 the supplier, the off sales, the lost sales, and the sell-off rate. The sell-off rate is the amount of product sold from the total amount purchased as a percentage. The store inventory is simply the stock level read from the inventory screen of the external file. For each store, the daily stock level is read from the screen into the stock variable. The consumer demand is the demand read from the demand screen of the external file. For each 20 store, the daily demand is read from the screen into the demand variable. The lost sales are calculated as the difference between the consumer demand and the actual sales. Both the consumer demand and the actual sales data are read from the external file. For each store, the lost sales, as calculated above, is read into the shortfall variable. The number of day's cover is calculated using both the stock level and the consumer 25 demand. The average daily demand over the next 5 weeks is calculated. Sum Demand Average Daily Demand = over next 5 weeks. No.of Days WO 03/054756 PCT/IB02/05580 79 The number of day's cover is calculated by dividing the stock level by the average daily demand. Stock Level Days Cover = Stock Level Average Daily Demand The code used to calculate the above data is held in the Actual Results module. 5 Once the data has been obtained, the data is output to the Results screen 730, shown in Fig. 22. The Supply Plot chart is activated and plots the data held on the Results screen. When the Level, Group or Store is changed the results are recalculated for the new Level, and the appropriate data is once again output to the Results screen. DEMAND GRAPH 10 The Demand Graph 710 only plots a graph of the consumer demand. It is accessed in a similar way to the Actual Graph. The Demand Graph 710 plots the current demand held by the model in a graphical format. The demand plotted may be either generated using actual sales or synthetic data. Once the demand has been loaded into the model the graph 710 may be plotted. This graph 710 is accessed via the Plot Base Demand tab on the main menu 300. On selecting 15 this tab, the model asks whether an external should be used. To plot the Demand Graph 710, the No button should be pressed. The demand data currently in memory will be written to the Results screen and the Demand Plot chart activated. As with the other graphs it may be plotted at different levels. When the Level, Group or Store is changed the results are recalculated for the new object, and the appropriate data is once again output to the Results screen. This is 20 performed in the Plotting module, and the Update Demand Plot procedure. SUPPLY GRAPH The supply graph may be a main graph used by the model. The results of the simulation are output here. The following variables may all be plotted in this graph.

WO 03/054756 PCT/IB02/05580 80 Description Example Variable Store Inventory Stock (Store, Day) Dispatch to Store SDispatch (Store, Day) DC Inventory DCStock (Day) Dispatch to DC DCDispatch (Day) Consumer Demand Demand (Store, Day) Lost Sales Shortfall (Store, Day) Days Cover StoreDays_Cover (Store, Day) During the simulation, the above arrays have data inserted into them. The stock level at each store is monitored throughout the course of the simulation; any dispatches to stores are also recorded, including the size and day of the dispatch. The same is also true for the distribution center. The consumer demand has already been loaded into the model and its data simply output 5 to the Results screen. If a store is out of stock, and there is some consumer demand, then a lost sale has occurred. Lost sales are also recorded throughout the course of the simulation. The number of day's cover is calculated once the simulation is complete. The graph should only be plotted once a simulation has been completed. When the graph is plotted data from the variables above is written. The Supply Plot chart is then activated, and 10 displays the data. When the Level, Group or Store is changed the results are recalculated for the new Level, and the appropriate data is once again output to the Results screen. This is performed in the Plotting module, and the Update Demand Plot procedure. ARCHIVE For some analyses, it is necessary to keep a record of the results achieved by different 15 replenishment strategies, and the parameters used to set them up. The archive provides a method WO 03/054756 PCT/IB02/05580 81 of saving all relevant information. The Archive screen may be accessed from the main menu 300 via the Results Archive button. The archive screen contains a list of parameters in each column of row 2. Row 3 contains the template for the parameters. This row references important parameters that need to be recorded. It also records some of the results of the simulation, as 5 these results are output to the Results screen. By changing the references it is possible to change the parameters that are recorded. The archive copies the current content of row 3, which reference all the important parameters, and pastes the values held here to the next blank row. The number of the row that the data should be pasted to is held in Archive screen. 10 The Archive may be extended to include other parameters as necessary. To do this, the parameter name should be added to a column in row 2, where the data required is held. PUSH PULL ANALYSIS The push pull analysis screen records certain parameters that the archive analysis screen may not. 15 The push pull analysis screen outputs the following information: Parameter Description Columns Date Date simulation was run 41 Time Time Simulation was run 42 Runcode Runcode from archive 43 Flow Option Type of Flow 45 49 Store Group ID Identification name for Store Groups 59 Store Group Size Number of Stores in each store group 60 65 WO 03/054756 PCT/IB02/05580 82 Initial Purchase Amount purchased (only applies to push 72 Quantity options) Supplier Minimum that may be ordered from a 73 Minimum Order supplier. Quantity Supplier to DC 74 Lead Time DC Safety Stock Number of Weeks cover held by DC 75 DC to Store Lead 76 Time Number of 77 delivery days per week Total Demand Total Demand over the period 86 Av. Average Daily Demand in all stores 87 Peak Peak daily demand in all stores 88 SD Standard Deviation of daily demand in 89 all stores Total Sales Total Sales over the period 91 Av. Average Daily Sales in all stores 92 Peak Peak daily sales in all stores 93 SD Standard Deviation of daily sales in all 94 stores Av. Average Individual Store Inventory 96 Peak Peak Individual Store Inventory 97 SD Standard Deviation of Individual Store 98 Inventory WO 03/054756 PCT/IB02/05580 83 Inventory VDR2DC Number of Deliveries from the vendor 100 to the DC SGI - SG6 Number of deliveries from DC to Store 101 -106 by store group Avail Availability Levels achieved 108 Sell Off Percentage of goods sold 109 Off Sale Percentage of Time item was off sale 110 Lost Sale Lost sales expressed as a percentage of 111 demand Av Average DC Inventory 112 Peak Peak DC Inventory 113 SD Standard deviation in DC inventory 114 Push Quantity Amount pushed to each store 136-233 Freq. Frequency of Pulls, number per day 234 Av. Quantity Average Size of Pull 235 SD Standard deviation in size of pulls 236 Initial Initial Store Stock 237 Final Final Store Stock 238 Initial Initial DC Stock 239 Final Final DC stock 240 WO 03/054756 PCT/IBO2/05580 84 When the simulation is run, all of the above parameters are calculated, and the values output to the appropriate places in the Push Pull Analysis screen. All of the code controlling the Push Pull Analysis is held in the Output module, with the main procedure Output Control, calling all the other procedures. 5 The inventive method may be embodied as computer readable instructions stored on a computer readable medium such as a floppy disk 124, hard disk 118, or system memory 112. Fig. 26 illustrates a block diagram of a computer readable medium 801 that may be used in accordance with one or more of the above embodiments. The computer readable medium 801 stores computer executable components, or software modules, 803-813. More or fewer software 10 modules may alternatively be used. Each component may be an executable program, a data link library, a configuration file, a database, a graphical image, a binary data file, a text data file, an object file, a source code file, or the like. When one or more of the software modules are executed by processor 110, the software modules interact to cause the computer system 100 to perform according to one or more embodiments of the invention as taught herein. 15 Having described several embodiments of the system and method of optimizing a supply chain in accordance with the present invention, it is believed that other modifications, variations and changes will be suggested to those skilled in the art in view of the description set forth above. It is therefor to be understood that all such variations, modifications and changes are believed to fall within the scope of the invention as defined in the appended claims. 20

Claims (41)

1. 2. A method of optimizing a supply chain as recited in Claim 1, wherein said calculating step is performed from sales and inventory data from said supply chain data. 15 3. A method of optimizing a supply chain as recited in Claim 1, wherein said selecting of products representing segments of said supply chain is performed to provide a simplified picture of a complete product range in said supply chain.
2. 4. A method of optimizing a supply chain as recited in Claim 1, wherein said supply chain data is formatted into spreadsheet format. 20 5. A method of optimizing a supply chain as recited in Claim 1, wherein said step of developing a simulation of said supply chain comprises: providing graphical and numeric output.
3. 6. A method of optimizing a supply chain as recited in Claim 5, wherein said step of providing graphical and numeric output delineates actual sales, transportation movements, 25 product inventory levels, and missed sales opportunities.
4. 7. A method of optimizing a supply chain as recited in Claim 1, further comprising the step of grouping similar stores and similar distribution centers in said supply chain. WO 03/054756 PCT/IB02/05580 86
5. 8. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors is used to develop future product flow paths.
6. 9. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors is used to develop product development strategies. 5 10. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors includes adjusting at least one of replenishment methods, service levels, days cover, minimum presentation stock, and replenishment days.
7. 11. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors includes adjusting supply chain options including at least one of minimum store 10 order quantity, minimum supplier order quantity, supplier to distribution center lead-time, distribution center to store lead-time, and distribution center safety stock.
8. 12. A method of optimizing a supply chain as recited in Claim 1, further comprising the step of producing a performance index for said simulated supply chain.
9. 13. A method of optimizing a supply chain as recited in Claim 12, wherein said 15 performance index is based on a gross profit after distribution expense.
10. 14. A method of optimizing a supply chain as recited in Claim 13, wherein said performance index for said gross profit after distribution expense for said simulated supply chain is divided by a gross profit after distribution expense for baseline.
11. 15. A computer-readable medium having computer-executable instructions for 20 performing steps comprising: plotting inventory throughout a supply chain plotting stores, distribution centers, and suppliers forming said supply chain; determining a base line for products in said inventory; defining features of said supply chain to be forecast; 25 defining replenishment characteristics of said inventory; running a simulation of said supply chain; and plotting inventory in said simulation. WO 03/054756 PCT/IB02/05580 87
12. 16. A computer-readable medium having computer-executable instructions for performing steps comprising: calculating demand from the weekly sales and inventory in a supply chain; loading new demand with current store groups; 5 performing base case analysis to obtain a baseline for each representative product in said inventory; recording results from said baseline analysis step; plotting said current supply chain; defining supply availability characteristics; 10 defining a forecast profile and replenish characteristics; running a simulation; and plotting availability of said inventory.
13. 17. A computer-readable medium having computer-executable instructions for performing steps comprising: 15 selecting products representing segments of said supply chain; collecting supply chain data relating to said selected products; calculating demand for said selected products; developing a simulation of said supply chain to produce a baseline for said supply chain; and
14. 20. altering factors effecting said simulated supply chain to produce an optimized supply chain. 18. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, wherein said calculating step is performed from sales and inventory data from said supply chain data. 25 19. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, wherein said selecting of products representing segments of said supply chain is performed to provide a simplified picture of a complete product range in said supply chain. WO 03/054756 PCT/IB02/05580 88 20. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, wherein said supply chain data is formatted into spreadsheet format.
15. 21. A computer-readable medium having computer-executable instructions for 5 performing steps as recited in Claim 17, wherein said step of developing a simulation of said supply chain comprises: providing graphical and numeric output.
16. 22. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 21, wherein said step of providing graphical and numeric. 10 output delineates actual sales, transportation movements, product inventory levels, and missed sales opportunities.
17. 23. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, further comprising the step of grouping similar stores and similar distribution centers in said supply chain. 15 24. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, wherein said step of altering factors is used to develop future product flow paths.
18. 25. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, wherein said step of altering factors is used to develop 20 product development strategies.
19. 26. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, wherein said step of altering factors includes adjusting at least one of replenishment methods, service levels, days cover, minimum presentation stock, and replenishment days. 25 27. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, wherein said step of altering factors includes adjusting supply chain options including at least one of minimum store order quantity, minimum supplier order quantity, supplier to distribution center lead-time, distribution center to store lead-time, and distribution center safety stock. WO 03/054756 PCT/IBO2/05580 89
20. 28. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 17, further comprising the step of producing a performance index for said simulated supply chain.
21. 29. A computer-readable medium having computer-executable instructions for 5 performing steps as recited in Claim 28, wherein said performance index is based on a gross profit after distribution expense.
22. 30. A computer-readable medium having computer-executable instructions for performing steps as recited in Claim 29, wherein said performance index for said gross profit after distribution expense for said simulated supply chain is divided by a gross profit after 10 distribution expense for baseline.
23. 31. In a computer system having a graphical user interface including a display and a user interface selection device, a method of providing and selecting from a menu on the display, comprising the steps of: selecting products representing segments of said supply chain; 15 collecting supply chain data relating to said selected products; calculating demand for said selected products; developing a simulation of said supply chain to produce a baseline for said supply chain; and altering factors effecting said simulated supply chain to produce an optimized supply 20 chain.
24. 32. A method of providing and selecting from a menu on the display as recited in Claim 31, wherein said calculating step is performed from sales and inventory data from said supply chain data.
25. 33. A method of providing and selecting from a menu on the display as recited in 25 Claim 31, wherein said selecting of products representing segments of said supply chain is performed to provide a simplified picture of a complete product range in said supply chain.
26. 34. A method of providing and selecting from a menu on the display as recited in Claim 31, wherein said supply chain data is formatted into spreadsheet format. WO 03/054756 PCT/IB02/05580 90
27. 35. A method of providing and selecting from a menu on the display as recited in Claim 31, wherein said step of developing a simulation of said supply chain comprises: providing graphical and numeric output.
28. 36. A method of providing and selecting from a menu on the display as recited in 5 Claim 35, wherein said step of providing graphical and numeric output delineates actual sales, transportation movements, product inventory levels, and missed sales opportunities.
29. 37. A method of providing and selecting from a menu on the display as recited in Claim 31, further comprising the step of grouping similar stores and similar distribution centers in said supply chain. 10 38. A method of providing and selecting from a menu on the display as recited in Claim 31, wherein said step of altering factors is used to develop future product flow paths.
30. 39. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors is used to develop product development strategies.
31. 40. A method of providing and selecting from a menu on the display as recited in 15 Claim 31, wherein said step of altering factors includes adjusting at least one of replenishment methods, service levels, days cover, minimum presentation stock, and replenishment days.
32. 41. A method of providing and selecting from a menu on the display as recited in Claim 31, wherein said step of altering factors includes adjusting supply chain options including at least one of minimum store order quantity, minimum supplier order quantity, supplier to 20 distribution center lead-time, distribution center to store lead-time, and distribution center safety stock.
33. 42. A method of providing and selecting from a menu on the display as recited in Claim 31, further comprising the step of producing a performance index for said simulated supply chain. 25 43. A method of providing and selecting from a menu on the display as recited in Claim 42, wherein said performance index is based on a gross profit after distribution expense. WO 03/054756 PCT/IB02/05580 91
34. 44. A method of providing and selecting from a menu on the display as recited in Claim 43, wherein said performance index for said gross profit after distribution expense for said simulated supply chain is divided by a gross profit after distribution expense for baseline.
35. 45. A system for optimizing a supply chain comprising: 5 a database of information concerning said supply chain; means for selecting products sent through said supply chain as representative products; means for creating a model of said supply chain from said information in said database and said selected products; means for altering factors in said supply chain to produce revised supply chains; and 10 means for comparing said revised supply chains to said model to optimized supply chain.
36. 46. A system for optimizing a supply chain comprising, as recited in Claim 45, wherein said factors in said supply chain include the method of product replenishment.
37. 47. A system for optimizing a supply chain comprising, as recited in Claim 46, wherein said method of product replenishment is one of: single push, push in waves, push 15 controlled push, push pull, continuous pull, and pull - cross docked.
38. 48. A system for optimizing a supply chain comprising, as recited in Claim 46, wherein said method of product replenishment is based on one of a min-max algorithm, an Adelsten specific method; and a dynamic method.
39. 49. A method of optimizing a supply chain as recited in Claim 1, wherein said step of 20 altering factors is performed to produce efficient product replenishment.
40. 50. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors is performed to produce efficient product store assortment.
41. 51. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors is performed to produce efficient promotion 25 52. A method of optimizing a supply chain as recited in Claim 1, wherein said step of altering factors is performed to produce efficient product introduction.