EP1518197A2

EP1518197A2 - Method and system for simulating order processing processes, corresponding computer program product, and corresponding computer-readable storage medium

Info

Publication number: EP1518197A2
Application number: EP03738007A
Authority: EP
Inventors: Stephan Hase; Jan Hickmann; Ulrich Wendt; Axel Wagenitz
Original assignee: Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV; Volkswagen AG
Current assignee: Fraunhofer Gesellschaft zur Forderung der Angewandten Forschung eV; Volkswagen AG
Priority date: 2002-06-25
Filing date: 2003-06-06
Publication date: 2005-03-30
Also published as: WO2004001633A3; US20060010017A1; AU2003245924A1; AU2003245924A8; WO2004001633A2

Abstract

The invention relates to a method for simulating order processing processes used for producing a complex product, particularly a motor vehicle, and a simulation system, a corresponding computer program product, and a corresponding computer-readable storage medium. The inventive method comprises the following steps: a) requirement figures for at least one class of the product are entered into a data processing device for at least one predefined period of time; b) said requirement figures are automatically adjusted to predefined sets of data describing production capacities and/or (production-)supply capacities by means of a computer program that is installed on the data processing device; c) the requirement figures or portions thereof are assigned to production sites (factories); d) the production and/or supply for the production is simulated based on the assignment done in step c); e) the distribution paths are automatically determined and the distribution(s) of finished products from the factories to the delivery locations is/are simulated; f) at least some of the data generated in steps a) to e) is stored and/or output.

Description

Process and simulation system for the simulation of order processing processes as well as the corresponding computer program product and corresponding computer readable storage medium

The invention relates to a method and a simulation system for simulating order processing processes for the production of a complex product, in particular a motor vehicle, and a corresponding computer program product and a corresponding computer-readable storage medium with the features mentioned in the preambles of claims 1, 16 and 18 to 21.

For several years now, efforts have been made in various branches of industry, particularly in the automotive industry, to improve manufacturing processes. One way to achieve such an improvement is to use simulation models to plan or check systems and in particular material flows. For this purpose, a wide range of tools was created, which are now available as standard software.

In the meantime, the focus has been broadened to include potentials that arise from the combination of several sub-processes, while taking into account the rules that serve to control the processes. However, this approach leads to a complexity that cannot be resolved with conventional instruments and methods.

Conventional methods and tools for material flow simulation u. a. are no longer enough to achieve the necessary transparency in such networked processes. New tools are therefore required here.

The investigation of business processes in the automotive industry is particularly demanding in many ways. A major reason for this is that for many sub-processes, considering the product "vehicle" without mapping the properties (equipment) is not expedient, since these properties have a strong impact on the process flow. When examining business processes, these properties must also be taken into account in such a way that, on the one hand, the behavior of buyers in certain markets is simulated and a number of rules regarding the composition of the vehicles are also observed.

The need for such a detailed product illustration arises, among other things, from the fact that planning algorithms work on the basis of parts or equipment and thus only a "realistic" product representation allows conclusions to be drawn about the efficiency of the processes.

The design or evaluation of complex business processes is difficult to carry out with previous methods. The possibilities for simulating these processes are limited, since conventional simulations are often only carried out on very abstract product and process representations.

At least for the automotive industry - as already indicated - it can be shown that the abstraction of process or product details quickly leads to a poor quality of the statements.

For example, an improvement of business processes across the boundaries of previously weakly coupled individual areas often leads to the strengthening of this weak coupling in integrative planning processes. In this way, for example, stocks in buffers between areas can be minimized. In addition to the desired reduction in stocks, it can also be observed that reducing the size of buffers can lead to a process that is more critical of faults. Such impacts have so far not been simulatively investigated in advance and cannot be validly assessed.

The design of overarching processes, such as the order processing process, is fraught with high risks without supporting tools. A lack of transparency also almost inevitably leads to inefficient processes. To minimize these risks, a system is therefore required that enables a process planner from the first draft - right through to implementation - to examine the impact of the planned processes qualitatively and / or quantitatively and, based on this, to design alternative process sequences if necessary. In particular, the known tools do not allow the complicated interdependencies of the vehicle equipment to be depicted with one another, to stay with the example of the automotive industry. There are therefore no tools with which it would be possible to generate vehicles in a model in which these dependencies are correctly taken into account (buildable vehicles).

So far, it has not been possible to evaluate the implementation of the planning procedure for plant accounting, capacity control, fault management and so on and / or alternative planning procedures qualitatively and quantitatively.

To examine the impact of the planned processes qualitatively and / or quantitatively and, based on this, to design alternative process sequences if necessary.

In particular, the known tools do not allow the complicated interdependencies of the vehicle equipment to be depicted with one another, to stay with the example of the automotive industry. There are therefore no tools with which it would be possible to generate vehicles in a model in which these dependencies are correctly taken into account (buildable vehicles).

The invention is therefore based on the object of providing a method and a simulation system for simulating order processing processes for producing a complex product, in particular a motor vehicle, as well as a corresponding computer program product and a corresponding computer-readable storage medium, by means of which the above-mentioned disadvantages are eliminated and in particular Holistic modeling and simulation of all planning processes in the logistics chain is made possible.

This object is achieved according to the invention by the features in the characterizing part of claims 1, 16 and 18 to 21 in cooperation with the features in the preamble. Appropriate embodiments of the invention are contained in the subclaims. A particular advantage of the method according to the invention for simulating order processing processes for producing a complex product, in particular a motor vehicle, is the possibility of holistic modeling and simulation of all planning processes in the logistics chain by performing the following steps:

a) input of required numbers for at least one class of the product into a data processing device for at least a predefinable period,

b) automatic comparison of these demand figures with predeterminable data sets describing manufacturing and / or (manufacturing) supply capacities by means of a computer program installed on the data processing device,

c) automatic assignment of the demand figures or parts of the demand figures to production locations (plants),

d) execution of a simulation of production and / or delivery for production based on the assignment made in step c),

e) automatic determination of the distribution channels and execution of a simulation of the distribution (s) of finished products from the plants to the delivery locations,

f) storage and / or output of at least part of the data generated by steps a) to e).

A simulation system for simulating order processing processes for the production of a complex product, in particular a motor vehicle, advantageously comprises the modules “forecast”, “fixed orders”, settlement “,“ production ”and“ distribution ”, the modules being controlled by a computer System-implemented computer program, interact in such a way that the following steps can be carried out:

a) input of required numbers for at least one class of the product into a data processing device for at least a predefinable period, b) automatic comparison of these demand figures with predeterminable data sets describing manufacturing and / or (manufacturing) supply capacities by means of a computer program installed on the data processing device,

In a preferred embodiment of the simulation system, it is also provided that the simulation system has interfaces to databases of real systems, such as databases from dealers and / or production sites.

A computer program product for simulating order processing processes for the production of a complex product, in particular a motor vehicle, comprises a computer-readable storage medium on which a program is stored which enables a computer, after it has been loaded into the memory of the computer, a method for simulating To carry out order processing processes, the simulation comprising the method steps according to one of claims 1 to 15.

In order to carry out a simulation of order processing processes for the production of a complex product, in particular a motor vehicle, a computer-readable storage medium is advantageously used, on which a program is stored, which enables a computer after it has been loaded into the memory of the computer, a method to simulate order processing processes, the simulation comprising the method steps according to one of claims 1 to 15.

The use of a method for simulating Order processing processes according to one of claims 1 to 15 or a simulation system according to one of claims 16 or 17 for determining planning data such as optimization potential, decision alternatives, key figures for delivery time or delivery reliability, utilization of means of transport, costs or the like.

In addition, it is an advantage for strategic, tactical and / or operational decisions, planning data such as optimization potential,

Decision alternatives, key figures for delivery times. or delivery reliability, utilization of means of transport, costs or the like, which have been made available by a method for simulating order processing processes according to one of claims 1 to 15 or by a simulation system according to one of claims 16 or 17.

In a preferred embodiment of the method according to the invention, it is provided that the data records used in the automatic comparison of the required numbers in step b) according to claim 1 include restrictions of the production sites and / or suppliers.

In a further embodiment of the method according to the invention, it is also provided that the demand figures are determined in step a) according to claim 1 by specifying a first demand forecast for a first forecast period and determining a second demand forecast for a second forecast period using stochastic methods from the first forecast and the demand numbers are determined according to predefinable algorithms which evaluate the first and / or second demand forecast.

Furthermore, it proves to be advantageous if the automatic comparison in step b) according to claim 1 includes a correction of the required numbers to adapt to the manufacturing and / or (manufacturing) supplier capacities.

Another preferred embodiment of the method according to the invention provides that method steps a) to c) according to claim 1 comprise the following steps:

Specification of provisional demand numbers (demand forecast) for a first forecast period, preferably for a sales year,

simulative generation of dealer orders for a second forecast period, preferably for three months, Evaluation of the preliminary demand figures and dealer orders and determination of an updated demand forecast for the second forecast period,

Comparison of the updated demand forecast for the second forecast period with the capacities of the production facilities and / or the suppliers and determination of approved fixed order quotas and / or module quotas,

- Generation of the required numbers (settlements) for the predefinable period, preferably a delivery week, by evaluating the approved fixed order quotas, module quotas and / or of simulated buyer orders newly received by the dealers,

- Comparison of these demand figures (settlement) with restrictions (capacity,

Utilization or the like) of the manufacturing facility (s) and / or suppliers and allocation of the required numbers (settlement) to the manufacturing facility (s).

A further advantage of the method according to the invention is that the automatic allocation of the demand numbers to the production sites results in a division of the demand numbers of the specified period into daily settlements, or that the automatic assignment of the demand numbers to the manufacturing sites is the compilation of daily programs for the manufacturing sites or the dissolution of the products specified in the daily subscriptions in their modules.

It is also advantageous if the demand figures include information on the essential features of the products ("heavy items").

In a further preferred embodiment of the method according to the invention, it is provided that the model on which the simulation is based depicts a number of manufacturing locations, in the model on which the simulation is based include the parameters which characterize a production facility, the limit capacities, working time models, personnel base and / or in the model on which the simulation is based a differentiation is made between dealers, in particular between dealers on the domestic market and importers, and / or in the model on which the simulation is based, distribution channels are subdivided into sub-distribution channels. Another advantage of the invention is that the data generated by steps a) to e) according to claim 1, quantitative evaluations of process designs, evaluations of strategies such as fault management, freezing times of orders, delivery times, delivery reliability, means of transport utilization and / or costs include.

In addition, it is also advantageous if data from databases of real systems, in particular from databases of dealers and / or production sites, are automatically evaluated during the method.

The benefits associated with the use of the simulation model according to the invention during planning and operation can be summarized as follows:

- Validation of the planned order processing process prior to implementation,

Improving the planning of the order processing process in the run-up to implementation as well as in the operating phase by setting up and evaluating various scenarios in an experimental field

Analysis and evaluation of potential weak points, such as answering such exemplary questions as: At which points do unnecessarily long processing times occur and which boundary conditions are responsible for this?

Support of specification specification for the development of planning and control instruments,

Testing the scope for decisions and performance limits in extreme situations (overloads, disturbances etc.) and deriving possible compensation strategies as

Preventive measure.

The use of the method or simulation system according to the invention also offers the following advantages:

1. In the invention, a product model is included that allows the complex dependencies of the vehicle equipment to be represented by rules. This enables vehicles to be created in the model that meet these rules (buildable vehicles).

2. The detailed implementation of planning procedures for plant booking, capacity control, fault management and so on is made possible. alternative

Planning processes can be assessed qualitatively and quantitatively.

3. Almost any section of the OTD process (OTD = Order to Delivery) can be modeled, modified and the effects on the overall process can be examined.

4. Operative systems can be integrated so that the effects of decisions based on actual data can be examined in advance.

The method according to the invention integrates and extends concepts from material flow simulation, business process simulation and systems of supply chain management (SCM). The method according to the invention is based on the event-controlled, discrete simulation of business processes with a high level of detail in relation to product resolution, planning algorithms and mapping area. This result can also be used to advantage in logistics.

The invention simulates essential planning processes for the product representation and the order processing process that affect the entire logistics chain. The high quality of statements that can be achieved by the invention _/ should be emphasized as a particular advantage; this quality makes it possible to support the introduction of new processes on a strategic, tactical and operational level. This achieves a new holistic modeling and simulation of all planning processes in the logistics chain and enables the construction of complex models.

Only the high quality of the statements allows the invention to be used consistently from the process design through to operational use:

at a strategic level: process designs can be assessed quantitatively.

on a tactical level: evaluation of strategies, for example for fault management, is made possible. at the operational level: planners can evaluate production programs in terms of effects on delivery times, delivery reliability, utilization of means of transport or ultimately costs.

In addition to ensuring process design, there are a number of other advantages:

System loads can be generated efficiently based on complex regulations.

- Buffers between process stages can be dimensioned based on verified knowledge. ,

The order processing process can be examined both as a whole and in its sub-aspects.

Planning algorithms (demand and capacity management, disruption management, procurement and so on) can be examined and evaluated.

The formulation of planning procedures is clear. The simulation model can serve as a reference.

The process design is supported seamlessly from the first draft through the implementation of IT systems to operational operations.

The holistic evaluation of processes made possible by the invention enables new optimization potentials to be addressed and implemented with very low risks for the company.

Further preferred embodiments of the invention result from the other features mentioned in the subclaims.

The invention is described below in exemplary embodiments on the basis of the associated. Drawings explained in more detail. Show it:

Figure 1 is a flow diagram of a simulation study;

FIG. 2 shows a verification and validation of simulation models; Figure 3 is a comparison of a sequential and a simultaneous

Process architecture;

Figure 4 is an illustration of possible process stages in the implementation of

Day orientation;

Figure 5 shows an example of a possible structure of a model for simulating the

Order fulfillment process;

FIG. 6 shows an exemplary representation of levels of the vehicle description;

Figure 7 is a representation of input and output data of a simulation study for the

Model "Vehicle delivery at a delivery point of the manufacturer or manufacturers";

FIG. 8 shows a distribution of the production throughput times for the vehicle types X and

Y;

FIG. 9 shows a diagram for the real production lead time for vehicle type Y;

FIG. 10 shows a diagram for the real production lead time for vehicle type X;

FIG. 11 shows a representation of the reduction in stochastics in the production lead time of vehicle type Y;

Figure 12 is an illustration of the reduction in stochastics in the

Manufacturing lead time of vehicle type X;

FIG. 13 shows a comparative representation of the customer arrival time for a scenario in which the customer is only informed about the delivery date when the vehicle is finished (top) and a scenario in which the customer is already informed about the delivery date in the vehicle planning (bottom);

FIG. 14 shows a table of the space required when delivering the vehicle to a vehicle

Delivery point of the manufacturer or manufacturers; FIG. 15 shows a table of the results of the sensitivity analysis;

FIG. 16 shows a table of the space requirement for scenario S2 and the distribution Vc of the production cycle time;

Figure 17 is a table of the potential savings at parking spaces in the transition from

Scenario S1 to scenario S2;

Figure 18 is a table with information on reducing the capital commitment costs per

Year in case of transition from scenario S1 with distribution Va of the production lead time to scenario S2 with distribution Vc of the production lead time;

Figure 19 is a table with information on reducing the capital commitment costs per

Vehicle in the event of the transition from scenario S1 with distribution Va of the production lead time to scenario S2 with distribution Vc of the production lead time;

FIG. 20 shows a schematic representation of an annual forecast;

FIG. 21 shows a schematic representation of the forecast update;

FIG. 22 shows a schematic illustration of the capacity adjustment;

FIG. 23 shows a schematic illustration of the generation of “approved

Firm orders;

FIG. 24 shows a schematic representation of the “weekly settlement”;

FIG. 25 shows a schematic representation of the “daily settlements”;

FIG. 26a shows a composition of the US and Canada markets in one exemplary embodiment;

FIG. 26b shows the markets and dealers in one exemplary embodiment; Figure 27a PR number families for engines, transmissions and air conditioning;

Figure 27b PR number families for radios, paints and hoods;

FIG. 28 shows a product tree;

FIG. 29 shows a diagram to illustrate the sales forecast for the vehicle

X in one embodiment;

FIG. 30 is the basis of the diagram shown in FIG. 29

Sales data;

FIG. 31a shows a course of the proportional vehicle sales for North America and Europe used in one exemplary embodiment;

FIG. 31b the sales data on which the diagram shown in FIG. 31a is based;

FIG. 32a shows a course of the proportionate vehicle sales used in one exemplary embodiment for the USA, Canada and the “other regions”;

FIG. 32b the sales data on which the diagram shown in FIG. 31a is based;

Figure 33 is a table of days off in plant A in one

Embodiment;

FIG. 34 shows a probability distribution for the swirl at

Daily program formation in one embodiment;

Figures results of the simulation using an exemplary

35 to 42 basic models:

Figure 35 average, minimum and maximum delivery times; Figure 36 average delivery and order lead time;

Figure 37 delivery time for vehicles with a specific engine;

Figure 38 Scheduling accuracy; Figure 39 Weekly program reliability;

Figure 40 ZP8 fidelity;

Figure 41 delivery reliability;

Figure 42 stocks;

Figures results of the simulation using an exemplary

43 to 49 basic model and the additional assumption that a strike occurs that countermeasures are taken:

Figure 43 average delivery and order throughput time;

Figure 44 Delivery time for vehicles with a specific engine;

Figure 45 Scheduling accuracy;

Figure 46 Weekly program adherence;

Figure 47 ZP8 fidelity;

Figure 48 Delivery reliability;

Figure 49 stocks;

Figures results of the simulation using an exemplary

50 to 56 basic model and the additional assumption that a strike occurs and as a countermeasure the target output of the vehicles is reduced:

Figure 50 average delivery and order lead time;

Figure 51 delivery time for vehicles with a specific engine;

Figure 52 Scheduling accuracy; Figure 53 Weekly program adherence;

Figure 54 ZP8 fidelity;

Figure 55 Delivery reliability;

Figure 56 stocks.

In the following, the invention will be described in detail using various degrees of detail in order processing in the automotive industry.

Customers' demands on vehicles and on the service package that the automobile manufacturer attaches to its products are extremely diverse. In addition to optimal product quality, maximum economy and high operational reliability, the service, which does not begin with the after-sales service, but already with the vehicle purchase, plays a decisive role for the buying behavior of the customers. In particular, in order to optimize the services for customers prior to the delivery of new vehicles, the automotive industry aims to reduce delivery times as much as possible between customer order and vehicle handover, and to maximize delivery reliability.

It has been shown that in order to achieve this goal, it is generally necessary to redesign the business process of order processing, from the creation of annual sales forecasts and corresponding program planning to the triggering of customer orders and vehicle handover to the customer.

The task associated with the design of the order processing process results from the complexity of the system under consideration, in which the process runs, as well as the various options for consciously influencing system behavior, for example by changing the control principles or unconsciously due to disruptions.

In order to be able to assess and optimize the dynamic behavior of the system as well as the quality of the business process of order processing in the run-up to the implementation, and later, an instrument must be available as an experimental environment with the help of which quantified statements about the system behavior with varying system loads and alternative control principles can be provided become possible.

The following success parameters are particularly important for evaluating the quality of the order processing process:

Reduction of delivery times,

Ensuring delivery reliability,

Optimization of capacity utilization,

Minimization of vehicle stocks in the distribution system.

Flexibility ranges must be integrated into the order processing process, which enable quick adjustments to varying boundary conditions. Knowledge of these ranges and the room for reaction associated with them represent an important prerequisite for being able to combine the outlined success parameters to an overall optimum. Specifically, quantified statements are required about: the capacity limits of the various systems and subsystems (dealer network, sales, production, distribution),

the potential bottleneck capacities,

the effect of different control principles under varying boundary conditions,

the dependencies between material stocks, stocks of new cars, lead times, delivery dates, delivery reliability and capacity utilization,

the existing restrictions and their effects on the order processing process as well as the knowledge of possibilities to be able to remove restrictions in the short or long term.

In order to be able to assess and evaluate the dynamic behavior of the “vehicle production” system under a newly defined process organization that is described by the process of order processing, one or more simulations of the order processing process are carried out with the method according to the invention.

In order to achieve the goals specified below, a modular simulation model was set up, with the help of which the system behavior can be assessed under various boundary conditions. At the same time, the model shows optimization potential in order to be able to continuously improve the business process of order processing in the planning and implementation stages. In addition to this process modeling, weaknesses (e.g. bottleneck capacities) of the system can be identified and evaluated through simulation studies. The benefits of, for example, expanding capacity at a specific site can be quantified with regard to the success factors and possible investment decisions can be secured.

The exemplary simulation model shows, in addition to the dealer network of the Federal Republic, the central sales and planning areas, selected plant locations and the distribution of the manufacturer in an abstract form. In addition, the core processes of the suppliers, important order modules and the interfaces to the production areas of the automobile plants are roughly mapped. By using the exemplary simulation instrument, the following goals are achieved:

Integration of all manufacturers (group) and production sites working in the network in the process of order processing:

To do this, it is necessary to develop the order processing process as an open framework in order to be able to adapt the process to the brand, location and buyer-specific requirements. The necessary changes must be implemented in a simulation model and evaluated with regard to the effects on dynamic behavior.

The simulation model accordingly represents, on the one hand, a modeling environment for the adaptation of the open reference model, and, on the other hand, it represents a model memory for those already configured based on location and brand

Model components.

Evaluation of quantity deviations in long-term to short-term planning:

- The lead time, inventory and adherence to delivery dates largely depend on a balanced relationship between the available capacity and the required capacity. This comparison must already be carried out in the annual production program planning based on sales forecasts in order to align the company to the seasonal fluctuations (so-called breathing company). Long-term capacity planning must be coordinated with personnel planning in particular in order to take annual working time models into account.

With the help of the exemplary simulation, different annual working time models but also different shift models can be evaluated with regard to their effects on the breathing company. The flexibility ranges associated with the alternative methods of personnel planning can be assessed.

In addition to the possibilities of long-term capacity adjustments, the exemplary simulation model can also be used to test various methods of medium and short-term capacity matching in order to obtain statements such as the

System components (production, distribution, etc.) Fluctuations in capacity requirements react and which short-term control interventions with which result are possible (effects on lead times, delivery reliability, etc.).

The short-term fluctuations in capacity requirements are determined by volume-related

(Number of vehicles) and property-related deviations from the order

(Equipment variants) triggered.

In order to optimally coordinate the long-term to short-term quantity and capacity planning, the concept of the exemplary simulation tool

Planning functions have been integrated that require interaction between the dealers and the manufacturer. These planning functions include the creation of an annual sales forecast, the coordination of a sales target matched to the available capacities, the triggering of firm orders by the dealers as well as the implementation of settlements (conversion of firm orders into individual orders, the

Specify vehicles completely) and the permanent adjustment of the settings to customer orders or dealer specifications for vehicles without a specific customer order until shortly before the start of assembly.

The benefit for the dealer is in particular the possibility of existing ones

Adjust orders to customer orders up to three days before MO and thus meet customer requests at short notice.

Through the concept, the manufacturer gains greater planning security (guaranteed production volume through firm orders three months before MO), shorter

Delivery times and higher delivery reliability, which increases the competitiveness of the different brands.

Analysis and systematization of the restrictions of the order processing process:

Different restrictions can be provided when optimizing the business process. These include corporate strategies (diesel quotas etc.), collective agreements, supplier capacities, plant restrictions etc. The effects of these restrictions on the success factors of the order processing process can be assessed with the simulation tool in order to

Receive decision support, which advantages are associated with the removal of individual restrictions. Analysis of performance limits and assessment of measures to push the limits:

- The performance limits of individual systems are determined by the technical and personnel capacities. Experiments in the exemplary simulation model with different system loads can make potential bottleneck capacities transparent under various boundary conditions (for example, serious deviations in ordering behavior from the forecast values).

In particular, the possibilities and benefits of a flexible distribution of the production program to different plant locations for individual models and types are made assessable in terms of capacity utilization, delivery times and other success factors through simulation studies.

Evaluation and optimization of the supplier interface in the order processing process:

- The integration of the suppliers in the process of order processing plays a crucial role in order to avoid shortages, which inevitably extend delivery times and jeopardize adherence to delivery dates. With the help of the simulation model can be checked. what information must be provided to the suppliers at what times, so that the delivery of important order modules (heavy items) can be realized as required.

With the simulation tool, it is possible to evaluate the usefulness of the information that the suppliers can derive from the sales forecast, the content of the firm orders and the changes in the settlement in order to optimize their production and distribution processes.

Evaluation of alternative distribution strategies:

With the help of the simulation, it can also be checked to what extent the objective of the distribution can be taken into account in the sequence planning in production and what advantages or disadvantages are associated with it. In a simulation, the dynamic process of order processing is simulated in a model in order to arrive at knowledge that can be transferred to reality. In general, simulation is a tool for dealing with reality.

Simulation is therefore an auxiliary tool to generate an image of the dynamic behavior of systems of reality in a model. As a system, the totality of elements with relationships between these elements and their properties is understood. Properties describing the system are its state variables. Changes in the state variables reflect changes in the system environment. Systems to be examined can be both facilities, such as factories, and processes. Specifically, it is about the investigation of system behavior in the event of changing environmental conditions. The areas of application for simulation are in particular where it is difficult to model complex, real world developments. The aim is to understand this complexity and consequently to master it. Uncertainties regarding the behavior of dynamic systems under changing environmental conditions can be eliminated more quickly and easily using simulation.

The complexity of real dynamic systems makes it difficult to determine the causal relationships between the influencing factors. Modeling as part of a simulation study enables a reduction in complexity by means of comparative-static analysis, which means that when comparing alternative process configurations, only one parameter is ever varied. The quantity of influence on the object under investigation is successively measured by the influencing parameters considered. The functionality of simulation instruments proves to be advantageous in viewing and evaluating the degree of influence in isolation. Furthermore, an improvement can result from the variation of influences that affect the system in parallel.

When modeling and simulating, for example, the order processing process of a car manufacturer, key indicators for evaluating the process, such as delivery time and delivery reliability, can be examined. In this way, potential for optimization can be identified and alternative decisions can be discussed. Simulation as a tool for mapping dynamic systems is used in different areas of the working world. Economic, sociological and ecological issues can be examined with the help of simulation.

The following areas are the focus when using simulators:

in science to investigate system behavior, in the technical area in system development, in system management, - in development planning.

The reasons for which a simulation is generally used are in the special case of order processing in automobile manufacture. Simulation is particularly useful when

We are writing in new territory, the limits of analytical methods have been reached, complex interrelationships are beyond our imagination, experimenting with the real system is not possible or is too cost-intensive, and the time behavior of a system is to be examined.

The use of a simulation tool when examining the order processing process - as in production and logistics - aims to achieve the following special goals, among others:

Improvement of planning quality and planning security, generation of an understanding of the system and mastery of system complexity.

From a business perspective, simulation technology is of great importance as an auxiliary tool in the decision-making of such complex processes. Decision alternatives can be simulated in advance. Simulation results provide information about the consequences of the respective decision in relation to the entire system. Possible wrong decisions can thus be avoided ex ante. In summary, simulation tools are usually used where relatively reliable statements about the behavior of complex, dynamic systems can be made with comparatively little effort.

The use of simulation tools has therefore proven to be particularly advantageous in the following areas:

system understanding

- parameter sensitivities,

Justifiability and verifiability of the chosen solution, avoidance and elimination of bottlenecks, dynamic analysis and presentation of the entire process.

Inexpensive solution

Optimization of processes,

Saving or simplification of system elements,

Optimization of buffer sizes and stocks.

Increased safety

Confirmation of the planning project, minimization of the entrepreneurial risk, - functionality of the planned system,

Functionality of the control.

So-called reference models were used for the simulation tool under consideration. These reference models contain, for example, prefabricated components. Examples of building blocks are machines, means of transport or warehouses. This considerably reduces the effort required to carry out the actual simulation experiments.

The building blocks provided can be combined depending on the particular question. The modeling effort of the user is then limited to the parameterization of the functionalities of the blocks used. With the help of this principle, model creation can be significantly accelerated. Especially in the case of very complex issues, it is necessary to abstract from reality in the context of modeling. A detailed reproduction of the real system is not possible in most cases. In this context, the significance of the model regarding reality based on the chosen level of abstraction must be checked. The latter has a decisive influence on the assessment of the simulation results and thus the benefits of the simulation method in general.

The definition of the degree of abstraction in the modeling is part of the simulation study, which will be briefly described below.

Figure 1 shows schematically the sequence when carrying out a simulation study.

In contrast to representations of simplified processes, the scheme shown in FIG. 1 takes into account the problem that a model may have to be adapted after verification and validation.

The individual components of a simulation study are briefly explained below.

As part of a problem definition, it must be clarified which question is to be investigated using simulation. This aspect is particularly important for the delimitation of relevant and irrelevant areas of the real system to be represented in the model. Depending on the question, it must be determined which influencing parameters of the system must be taken into account in the model. After clarifying this fact, it can be determined what level of detail is required when creating the model in order to be able to make qualified statements regarding the problem. Depending on the abstraction, the effort for modeling can be derived.

In order to represent the problem defined in the first step in a suitable manner, the relevant input variables have to be systematized. It is also necessary to discuss which parameters significantly influence the system. The relevant influencing parameters are then classified. In the context of system analysis, the dynamic parameters are of particular interest. The determination of the quantity of system influence of an individual parameter is usually supported by a sensitivity analysis. Dependencies between the input and output variables must be taken into account as a characteristic of the system behavior when developing the model. The input data required is collected in parallel with the model development.

After the development of the formal model, the contents of the model concept must be transferred to the computer program. During the model conception, the system components are worked out, which must be mapped in the model. These are suitable and can be transferred to the model in the required level of detail. Some simple structured questions can be used to implement standard simulation tools. As mentioned above, many of these standardized simulation tools are based on the so-called building block principle. Tools of this type can be used relatively flexibly in some fields of application, but generally prove to be unusable in the case of complex problems, such as for example in order processing in the automotive industry. It was also necessary to recreate the corresponding software for the exemplary simulation tool.

Following the model implementation, both verification and validation of the model must be carried out. Verification is the review of the individual steps in the modeling process. In other words, it is checked whether the model relationships formally developed as part of the model concept have actually been implemented in the computer model.

In contrast, the validation of the model includes the question of whether reality, based on the objective of the simulation, is appropriately and correctly represented in the model. Validation is simple in cases where a real existing system has to be modeled. First, the state of the real system has to be shown in a model. The model is suitable if the simulation results match those of the real system provided that comparable parameter settings are made. The model must map the behavior of the real system precisely enough and without errors.

In simple terms, verification means checking whether the right things are mapped in the model, while validation checks whether things are mapped correctly.

In the event that doubts about the correctness of the model are justified after the verification and / or validation, the simulation study must be carried out at this stage to be interrupted. At this point, the feedback for model development takes place. Figure 2 illustrates the relationship between model verification and model validation.

After the quality check has been successfully completed, the experiments can begin.

There are several aspects to consider when evaluating the results of the simulation runs. The following statistical principles must be observed:

- If dynamic processes are simulated, the modeled system needs a certain time of "settling" until the stability of the model is given (the values from the phase of the so-called "warming up" period are to be excluded from the overall evaluation).

- Characteristic of simulation experiments with stochastic

Distribution functions is the independence of the simulation runs.

In order to increase the accuracy of the results of simulation runs with stochastic distribution functions, n replications of simulation experiments are carried out with the same parameter setting and an average value over n is determined.

Another method is to set a confidence or trust interval. This interval limits the range in which the true value to be determined lies with a certain probability.

After the simulation runs have been carried out, the results must be interpreted. Based on the simulation experiments, the user can gain insights into the behavior of the real system. However, the interpretation should always be made against the background of the abstraction made.

Simulation experiments carried out in a company usually serve as decision aids for management decisions. The results of operationally used simulation tools in particular influence real decisions. In this context, the quality of the results is of course an essential prerequisite for user acceptance. The modeling for the specific example of the simulation of an order processing process for the production of motor vehicles is to be explained in detail below.

The order processing process comprises all sub-processes from the customer order to the delivery of the vehicle to the customer. The following steps can be specified as subprocesses of order processing:

Order acceptance: Dealers and customers agree on the vehicle type, equipment and delivery date. If the customer orders the vehicle under the agreed conditions, the order is forwarded to the manufacturer's sales department.

Program planning: Customer and dealer orders are scheduled taking existing restrictions into account.

Production: After an order has been assigned to a plant and scheduled in a weekly or daily program, the vehicles are manufactured according to the production program.

Distribution: After completion and final acceptance of the vehicle in the factory, the vehicle is loaded and transported to an interim storage facility or directly to the appropriate dealer. The vehicle is then handed over to the dealer.

The market demand is forecast beforehand. Based on this forecast, the quantity planning is carried out after checking the available capacities. This planning is done for both vehicles and properties. The parts requirement can be determined on the basis of the quantity planning. Quantity planning also defines the ranges for the scheduling of orders. The aim of optimizing order processing is to significantly reduce delivery times. This is achieved through the use of the invention.

A solution for the optimization of order processing is the implementation of a new process structure. One example is the substitution of a sequential one

Process architecture through a simultaneous process architecture a possibility of Redesign of the process structure. The individual chain links - forecasting, program planning, manufacturing, and distribution - systematically interlock.

A key aspect of this effort is the introduction of a process in which a vehicle in order control can be assigned to a specific customer as late as possible.

Figure 3 compares the two process structures.

In order to achieve this reduction in delivery times, efforts are being made, for example, to switch vehicle scheduling in the factories from a ZP8 week to a ZP8 day (ZP8 = point of delivery 8; delivery point 8 marks the completion of a vehicle). In contrast to today's processes, with the daily orientation instead of the order backlog of a calendar week, only the order backlog of one day is always transferred to production. This process is called "day orientation". In contrast to the classic order processing, the process "day orientation" can shift the time of freezing (EZP) of the order. The freezing point marks the latest permitted date for changing a property or an order. The basic idea behind each of the process stages shown in FIG. 4 is to increase the flexibility in order processing.

Depending on the property to be changed, the specification of the order may be changed until shortly before production begins.

The introduction of the "daily orientation" process requires certain prerequisites to be met. For example, the stability of the release of operating dates must be guaranteed. Furthermore, the lead times for the procurement of the parts must not be outside the time window for changing the specification of the order the change options are limited to a small scope of properties, otherwise the advantage resulting from the flexibility of the order processing would not be exhausted and the delivery times could not be significantly reduced.

The advantage of the daily orientation compared to the "2 + 2" process described below is that an order is fixed in order control on a ZP8 day for undisturbed orders. Turbulence of orders in production, for example, as in "2 + 2 "Process, are then no longer possible. This measure results in a constant sequence of orders before production. The daily orientation stabilizes and thus simplifies the process. A consequent changeover to the modified process structure might cause unforeseeable complications. For this reason, various process stages have been developed that enable a smooth transition. These process stages are shown in FIG. 4.

The delivery time is also gradually reduced through the gradual, evolutionary introduction of the respective process level.

Since problems could occur as mentioned in the implementation of the "Daily orientation" process stage, a preliminary stage "2 + 2" is interposed in the exemplary procedure. Under certain conditions, this process already enables a lead time of fourteen calendar days. "2 + 2" is intended to clarify that with an order time of four (= 2 + 2) weeks by the retailer, order changes are possible up to approximately two weeks before the ZP8 date - in contrast to "1 + 3", where four are also (= 1 + 3) weeks between the order and the production date, but the order is frozen three weeks before the ZP8 date, so that an order can only be changed up to three weeks before the ZP8 date.

Compared to the "1 + 3", the throughput time is reduced by one week. The retailer thus orders a vehicle from "2 + 2" at least four weeks before the planned production date (ZP8 week). Order changes are then possible up to two weeks before the ZP8 appointment. In this case, changing an order does not lead to a postponement. The retailer can also place an order in the systems without a specific customer order. The property specifications of a dealer order can then be changed according to customer requirements up to the time of freezing.

As can be seen, the order processing process in the automotive industry is a very complex system with numerous subordinate system components and cross-connections, so that it is difficult to master the functioning of the individual chain links of the process in the context of an overall view of the real process. For this reason, the simulation tool described in more detail below is used. The order processing process is transferred to a model and simulated. Based on the efforts to optimize the overall process, changes in parameter settings can be used to determine the effects in all sections of the process chain. An idealized system state is first modeled in the example simulation. The result of the simulation of this reference model forms, so to speak, the benchmark for subsequent simulation runs, whereby the system status is now varied by the fact that events that can be observed in reality hinder the continuity of the process flow. In addition, so-called “worst case” studies can be carried out. This determines the constellation in which the system becomes unstable with regard to the parameter setting, that is to say the functionality of the system is endangered.

This form of system analysis is not practical on the real system, since this procedure would be very cost-intensive.

A complete model of order processing is divided into the sub-areas of vehicle construction, sales, planning, markets / distribution and plants. FIG. 5 shows an overview of the complete structure of a simulation model and the logical links between the system components. In the following, the above-mentioned components of the model structure - vehicle description, sales, process control, markets / distribution, factories - and their components are briefly explained in general and then explained in two more detailed execution stages. In parallel, the first case shows how the laws of the real system are implemented in the model structure. Following this, the procedure for modeling is explained using a concrete simulation study as an example.

paragraphs

Under the heading "Sales", a value trend for the global, absolute sales of the vehicle categories under consideration is to be specified in relation to the simulation period. Depending on the question, it is possible to define seasonal sales fluctuations by entering sales figures, for example, on a monthly basis instead of aggregated based of a year. This functionality is also important, for example, in connection with the key figure "breathing factory". vehicle description

For example, a vehicle is fully described by a six-digit model key. Components of this key include information about the vehicle class (platform and series), the designation of the body shape (sedan, variant and so on), the equipment level (base, Trendline, Comfortline, Highline and so on), as well as the designations for the engine and transmission.

Hierarchically subordinate to the model key, the detailed resolution of the vehicle takes place via a list of so-called "PR numbers". The unique description of a property is given via PR numbers. Each property is assigned to exactly one PR number. PR numbers are assigned to " PR number families "summarized. For example, the PR numbers "without airbag" and "airbag for driver" are assigned to the PR number family "airbag (short name AIB)". Each vehicle is uniquely described by exactly one PR number from each PR number family. This structure is also taken into account in the construction of the model.

Country or market specifics in the vehicle description are to be defined separately. Examples of this are vehicles with right-hand drive or special emissions standards.

The vehicle body takes different levels into account. This structure is shown in FIG. 6. In this example, the levels group (root, 1st level), platform (2nd level), vehicle type or class (3rd level), body shape (4th level), equipment (5th level) and country code (6th Level) is taken into account. The number of levels depends on the model scope. If, for example, only one platform or only one vehicle class of a platform is considered, the relevant levels can be combined accordingly. In the opposite case, the level of detail can be increased by modeling further levels. As part of the modeling process, it is to be determined ex ante to what extent the real structure of the vehicle levels can be abstracted.

Furthermore, installation rates (EWC) must be defined at all levels. The installation rate of one. Property defines the proportion of this property in relation to the family of properties. If only one vehicle description is modeled for each level, the EWC corresponds to 100% for each level. Other attributes at each level are "PR numbering", "PR numbering specifications" and "PR numbering groups". PR numbering means all those properties that are already installed in the basic equipment of the respective vehicle description, whereby one property of a PR number family is to be modeled as a setting. The combination of properties is called a PR number group. This functionality could be used to map constraints and prohibitions, for example. On the one hand, the need to form combinations of properties could be due to technical reasons. On the other hand, sales functions can also be mapped using this functionality. For this reason, in some cases the customer can only choose from a large number of so-called equipment packages, which reduces the number of variants and facilitates the property forecast.

PR number specifications, in turn, are all properties taken into account in the model for each family of properties, from which the customer can individually configure his vehicle. Installation rates must be defined for both the PR number groups and the PR number specifications. It is not necessary to enter an EWC for the properties set. The EWC of the property set in each case results from the difference of the sum of the EWC of the properties not set of a PR number family to one.

process control

As part of the process control, the time horizon for the forecasts and the planning of the vehicles is determined. The various process stages, as described above in connection with the changeover to a modified process structure, can be parameterized using appointment series. The rolling planning rhythm is mapped via the appointment series. For this, times and the corresponding time intervals of the events during the planning process must be defined.

Markets / Distribution

In the markets, a distinction can be made between dealers (domestic market) and importers. This differentiation can be useful because, in contrast to the domestic market, vehicle orders from export markets are not based on customer orders, but usually on forecasts. In this case, the vehicles are specified by the importer. A dealer / importer is informed about the name, the location of the Destination station, the distribution channels, the preferred plants and specific dealer planning parameters completely described. If necessary, the distribution channels can be subdivided into so-called sub-distribution channels. This functionality enables the nesting of distribution channels. For example, there is the option of mapping both alternative means of transport and different routes. In the present embodiment, each distribution channel is defined via the attributes “starting point”, “destination point”, “timetable” (= time interval between two transports), “transport capacity”, “transport” and “loading time”.

Information regarding the dates for volume agreements and vehicle orders of a dealer / importer are part of the specific dealer planning parameters. Another feature is the classification of customers by segmentation. In the exemplary model, customers are differentiated according to their preferences regarding the delivery time.

works

A plant is fully described by the name, the capacity, the throughput time, closures, the destination station and the vehicle classes installed in the plant. As

Locking is the temporary impairment of the availability of the planned capacity of one or more properties. The capacity of a plant is determined by

Multiplication of the output per hour and the weekly working time of the plant determined.

Statutory holidays and company holidays are listed separately in the example simulation tool and taken into account in the capacity planning.

A degree of flexibility over time can be defined for both throughput and weekly working hours.

In relation to the throughput time, the planned throughput time and the throughput time distribution are to be specified as value curves in the exemplary embodiment under consideration. The scheduling of the order in production takes place on the basis of the planning lead time. Depending on the variance of the throughput time distribution, there are corresponding effects on the ZP8 fidelity. An improvement in ZP8 fidelity can be achieved, for example, by reducing the stochastic influences.

Based on the results, statements about the stability of the order processing process as a whole can then be derived. In case that Only the stochastics in production represent the cause of the destability in order processing, deterministic production times under otherwise identical conditions enable the exact determination of the delivery date.

Property locks are also specified at the factories in the exemplary embodiment. The description of the block is based on its definition. The definition includes information regarding the warning time, the duration and the start of the block as well as the proportionate capacity of the property that is affected by the block. In the context of the "Blocking" functionality, a shift in the operating date could also be modeled. A shift in the operating date is referred to when a property cannot be installed due to a variety of problems at the original date (synonymous with "Operating date", "Start of Production" (SOP) is also used This is particularly problematic in cases in which the failure to meet the planned appointment becomes known very late and can no longer be taken into account in the scheduling. If an appointment is postponed, the blocked property is replaced by a property of the same PR Due to the high level of complexity and the existence of numerous constraints and prohibitions on property combinations, a sufficiently long lead time must be provided for this purpose, in this context it becomes clear what effects a late announcement of a postponement of the operating date has on stability t of the process and as a result may have to meet the planned ZP8 date.

An indication of the installed vehicle descriptions and the corresponding relative proportions of the vehicle descriptions in the entire production of the respective vehicle description is only relevant if more than one plant is taken into account in the model and vehicle descriptions are not produced exclusively in one plant. Otherwise, the proportion is always 100%.

The exemplary embodiment of the simulation of the order processing process allows a wide range of questions to be examined. The effects of strategic as well as operational-tactical decision alternatives can be simulated. , The following examples are intended to give an impression of the universal use of the invention. It is analyzed in all cases how the described key figures delivery time, delivery reliability, capacity utilization and stocks vary if certain factors negatively influence the continuity of the process. The effects of strategic decisions can be examined regarding

the implementation of a new process level, the complexity reduction (modularization in procurement and sales), - alternative procurement strategies (e.g. modular sourcing), alternative manufacturing concepts (customer order production vs. warehouse production), site planning, alternative distribution channels, long-term capacity planning (technical capacity ), - new dealer network structures.

With the simulation tool according to the invention, the effects of operational-tactical measures as a result of:

- postponement of the deployment date,

Bottlenecks at suppliers, unpredictable demand for vehicles! Properties, temporary capacity restrictions in the plants due to a machine failure and the like.

In the following, the general sequence of a simulation study using the model “vehicle delivery at a delivery point of the manufacturer or the manufacturers” is to be described.

In contrast to the conventional handover of new cars at a dealer, the customer is offered the option of handing over the new vehicle at a delivery point of the manufacturer or the manufacturers in the examined model. As before, vehicle orders are also accepted by local dealers.

The difficulty associated with delivering the vehicles to a delivery point of the manufacturer or the manufacturers is that the vehicles that are to be handed over to the customer at a specific point in time must also be available at that point. Specifically, this claim results in 100% delivery reliability. This requires binding planning of the customer-specific vehicles and stable process flows. As part of the When planning, the personal preferences of the customer with regard to the time of delivery must be taken into account.

Using the simulation tool according to the invention, it can be shown how the logistical claim regarding delivery reliability described above could be realized. Based on the status quo of the delivery, it is determined which system adjustments are to be made in order to notify the customer of a binding delivery date at the time the customer receives the order from the dealer. In order to be able to guarantee the delivery of vehicles to customers on the agreed date, the customer is currently only informed of the earliest possible date of delivery when the vehicle is completed.

Since the takeover by the customer cannot always be done at short notice, the vehicles must be stored in the meantime. A certain capacity at parking spaces is required for this. Costs arise in the form of capital commitment, fixed costs of

Construction of storage space and variable storage costs. As part of this

Problems can also be considered in monetary terms. The aim of such an analysis is to show what savings could result if the knowledge gained with the simulation tool according to the invention were incorporated into the real process design. Due to the many influencing factors that the

Destabilizing the process, it is difficult to precisely quantify the space requirements. In the event that several influences act on the system in parallel, a lead time distribution with a high variance results. As a result, the binding delivery date promised to the customer would not be kept.

The aim of the simulation is to compare alternative process configurations. First of all - as already explained above - a model has to be created that shows the relevant properties of the real system. Not all sub-processes of the order processing process are relevant in relation to the special circumstances to be examined here as an example. Accordingly, the abstraction can be made appropriately.

To investigate the problem posed, for example, it is only necessary to model Plant A. Furthermore, only the proportion of vehicles manufactured in Plant A that is intended for the domestic market needs to be taken into account. In addition to the restriction to Plant A and the German market, only the following levels of vehicle construction are considered in the model in the special embodiment: Manufacturer,

Platforms: AOO and A, vehicle classes, - market / country code (domestic / market Germany).

The scope of the properties recorded in the exemplary model is also limited to the property families that are regularly classified as critical. In this context, an important criterion is considered to be that the model approximates the complexity of the vehicle description in reality. The complexity depicted in the exemplary model relates to the properties installed in Plant A for the German market in the case of the thirteen property families considered as examples. Accordingly, a total of thirteen property families with, for example, 68 properties or PR numbers must be taken into account. As part of the exemplary modeling, the PR number specifications and PR number setting are also mapped.

The actual or target installation rates of the properties or vehicle classes are taken from the respective planning systems. Since only vehicle control is considered in this example, modeling suppliers is irrelevant for answering the question. The deadlines of process level "2 + 2" implemented in plant A are also implemented in the model. The distribution of the production lead time assumed in the model must correspond to the actual production time in plant A. For this purpose, for example, the net production lead times must be shown. Machine downtimes (for example the Weekends) are subtracted from the gross production lead times. Since the production lead times for the individual vehicle types have different distributions, the assembly lines are modeled separately. At this point, no abstraction from reality can be made.

For the real layout of the production systems for a first vehicle type X in plant A, a division of production into three independent segments has been assumed as an example. The production lead times in the respective segments have also been set independently of each other. In order not to unnecessarily increase the complexity of the model, this separation can be dispensed with. Instead, an average value over the production times of the three segments would have to be formed. For the production of the vehicle type X, a planning lead time for the production of approximately 75 h and for the production of a second vehicle type Y of approximately 60 h was assumed, whereby the planning lead time of the production results from the total of the planning lead times for the bodyshell, paint and assembly.

In the exemplary simulation, only logistical or production-related influencing factors should be evaluated in the process analysis. Figure 7 shows an overview of the input and output data required for this model.

Implement the model

General implementation of the model concept

The model concept is first implemented as a reference model in the simulation model. A period of two years was defined as the simulation period with a vehicle volume of around 500,000 vehicles. The distribution of the production lead time is shown in the model via a histogram, that is, intervals of lead times are determined and the percentage of vehicles whose production lead time has a value within the limits of these intervals is determined.

Figure 8 shows the real lead time distributions for the production of the two vehicle types X and Y in plant A.

The throughput time distributions for the two vehicle types X and Y are shown cumulatively in FIG. 9 and FIG. 10.

Configurations of two different process design scenarios are presented below for the simulation tool according to the invention.

In a first embodiment, the reference model in scenario S1 is modified by taking into account the special features that occur in connection with the vehicle delivery at a delivery point of the manufacturer or the manufacturer. As already described above, a binding earliest possible delivery date is only set when the vehicle is completed (ZP8). In the period between the completion of the vehicle and the handover of the vehicle to the customer at a delivery point of the manufacturer or the manufacturer, the vehicle must be temporarily stored in the designated parking spaces. Based on experience, an average service life of fourteen to sixteen calendar days was expected. However, this guideline can vary. A time-delaying element within the distribution channel was modeled in order to appropriately map the customer arrival time. The distribution referred to in the model as customer arrival time corresponds to the time span between the information of the customer about the earliest possible delivery of the vehicle and the actual time of delivery of the vehicle. The average customer arrival time is around sixteen calendar days. A capacity of around 8,700 parking spaces on the factory premises was provided for the temporary storage of the vehicles. In addition, alternative places of indefinite capacity can be taken into account.

In a first simulation step, the space required for the three expansion stages described is determined, the real distribution of the production lead time of the two vehicle types X and Y being assumed (distribution Va in FIGS. 11 and 12, respectively). Subsequently, the variance in the distribution of the production cycle time is successively reduced (distribution Vb and Vc in FIGS. 11 and 12, respectively). This measure increases ZP8 loyalty. As part of this sensitivity analysis, the influence of stochastics in the distribution of the production cycle time on the ZP8 fidelity is quantified.

FIG. 11 and FIG. 12 show the distributions Va, Vb and Vc of the production throughput time for the manufacture of vehicle type X and for the manufacture of vehicle type Y. In both cases, the course of the production throughput time distribution shows the reduction in the stochastic influences.

A further embodiment of the simulation tool according to the invention is described below, a second scenario S2 having been implemented in the configuration.

The scenario S2 of the second embodiment is based on the findings of the sensitivity analysis with regard to reducing the variance in the production lead time. A stable manufacturing process enables the timing of the transmission of the information of the earliest possible delivery date to the customer. As a result, the average service life of the vehicle is reduced according to ZP8. Optimally the delivery date preferred by the customer is already taken into account when planning the vehicle. In contrast to scenario S1, the distribution of the customer arrival time is not implemented in the distribution, but rather as a customer request date distribution in the order planning.

The two scenarios are compared in FIG. In order to be able to better compare both scenarios, exactly the same distribution was assumed in scenario S2 for the customer request date as for the customer arrival time in scenario S1.

After implementation, the model was verified and validated. As already described in the introduction, a comparison between the model and the model concept is carried out during the verification, i.e. it is checked whether all relevant system relationships that were defined in the concept are mapped in the model.

When checking the validity of the model, the simulation results were compared with the data from the real system. This procedure is recommended in cases where real input data are used. In the exemplary simulation, both sub-processes and the overall process were checked. For the validation of the overall process, the real fidelity measurement values from the fidelity measurement for plant A were used.

Depending on the question, it may be sufficient if the evaluation is based on only one simulation run. This approach is justified in connection with the present question, for example, because in a first step, only a rough estimate of the space requirement is to be made. In the event that exact values have to be determined, statistical methods must be used. This includes performing multiple replications, among other things. It must be ensured that the starting value is varied by the random generator when generating the random numbers. Otherwise the results are reproduced with each replication.

In addition, a rough estimate is usually sufficient to determine the duration of the "warming-up" period. This procedure is generally permissible since only a maximum value has to be determined for the specific question. However, approximation methods would be used, for example, if an average value is to be determined over a certain output variable. When evaluating the data from the exemplary simulation runs, a duration of three months was determined that the modeled system needs to settle.

FIG. 14 shows the results of the determination of the space requirement on the basis of the real distribution of the production lead time (distribution Va) for the expansion stage 1 (delivery of 300 vehicles per day), expansion stage 2 (600 vehicles per day) and expansion stage 3 (1,000 vehicles per day).

In FIG. 15, the results of the sensitivity analysis described above with regard to the effect of stochastic production throughput times on the ZP8 fidelity were processed. The values for the standard deviation in the production lead time shown in FIG. 15 were determined from a sample of 1,000 vehicle orders for each of the three assumed distributions. The expected value for the production lead time corresponds to the plan lead time.

As part of the further course of action, it can prove useful to assume that a transition to the process design of scenario S2 is only considered if the manufacturing process is stable. This prerequisite is given at least in approaches for the production lead time distribution Vc.

FIG. 16 shows the space requirement for the case of the distribution Vc of the production cycle time. In addition, Figure 17 shows the savings potential with regard to the space requirements compared to the current process in scenario S1.

The monetary analysis carried out in the context of the exemplary simulation resulted in the savings potential documented in FIG. 18, which in this example result exclusively from the reduction in the capital commitment costs. An investigation, which also looked at the cost of providing space for the temporary storage of the vehicles, resulted in a correspondingly higher amount.

The calculation of the potential savings when substituting the parameters from scenario S1 with the production cycle time distribution Va by the parameters of scenario S2 with the production cycle time distribution Vc is based on the following equation: annual savings potential of process changes

= Interest * average selling price * reduction of the average

inventory

An interest rate of 8% was assumed as the capital market rate. The average selling price of the vehicle classes considered in the simulation is 30,000 euros. The difference of the average inventory in both scenarios has already been determined in Figure 17. It should also be noted that only the vehicles manufactured in Plant A were taken into account in the calculation. In this case, too, the amounts would increase accordingly if the total vehicle volume delivered to the customer were included in the overall calculation.

In relation to a single vehicle, the average savings shown in FIG. 19 result depending on the respective expansion stage.

In a preferred embodiment, the simulation tool is divided into different program blocks, which implement the steps described above, which are required in the simulation for examining the respective problem. These program blocks have the following functional scope:

Program block "System load generator"

The system load generator generates the demand forecasts of the dealers and the orders of the buyers in a simple form. The forecasts are continuously (for example monthly) adjusted to the generated (actual) needs of the dealers.

The system load generator generates a simplified forecast of the number of cars that could be sold over the following year individually for the dealers and once in total at the beginning of a sales year. In addition to the number of vehicles that are likely to be required, their main features ("heavy items") are characterized.

The annual forecast can, for example, have the form shown in FIG. 20:

It should be emphasized that this is a simplified forecast that approximates the forecast progression of the retailers. Using the simulated annual process, the load generator generates purchase orders for each dealer, whose curve display has a similarly simplified form that deviates from the annual forecast.

Depending on the (middle) curve of the purchase orders of the previous months, the forecasts for the next period (e.g. 3 months) are updated per week.

This results, for example, in a curve as shown in FIG. 21.

The output data of the system load generator include the annual forecasts of the dealers, the buyers' requirements as well as the updated weekly requirements (number and "heavy items" of the required vehicles) of the dealers for the next forecast period. They simplify the input load for the following program blocks and can also be graphically illustrated become.

Program block "Capacity leveling"

The weekly requirements of the dealers are compared and adjusted in this program block with the abstracted modeled capacities of the plants and the roughly modeled capacities of the suppliers (see below).

For this purpose, the needs of the dealers are collected and accumulated. After a need correction by central sales, abstracted in the model, for example, represented by demand and capacity limits, the requirements (number, items) are compared with the actual capacities of the plants and the suppliers.

The capacity leveling can be represented as shown in FIG. 22.

The capacity leveling results in approved fixed order quotas for each planning week, which are then communicated to the dealers in line with the given module quotas (see Figure 23).

The output of the program block for capacity balancing includes the approved firm orders and module quotas for the dealers. They form the input for the program block "settlement generator". Program block settlement generator "

The settlement generator generates specific settlements (fixed orders and order modules) for each delivery week from each of the approved firm orders and module contingents (see Figure 24). The specification of the fixed orders for settlement depends on the purchaser orders received so far by the dealer and the requirements forecast for the delivery week. This specification is modeled in the model by means of probability distributions with mean values and scatter.

Settlement generator outputs are the settlements per dealer and delivery week. These individual settings are assigned to the plants by the following program block and converted into daily settings.

Program block "plant assignment"

The entries for the program block "Plant assignment" are the traders' settings. These are compared with the restrictions (capacities, capacity utilization, etc.) of the production plants and suppliers. In this program block, the plant assignment and the division of the weekly settlements into the delivery days take place (see Figure 25).

The output data of the program block "Plant assignment" include specific settings with information about the manufacturing plant, the suppliers involved and the name of the delivery day.

The settled concretizations are communicated to the dealers and form the input for the following program block.

Program block "settlement manipulator"

Throughout the process

Forecast -> Fixed Orders -> Settlements -> Daily Programs (see below) -> Production -> Distribution

meet buyer needs at the dealers. In the opposite direction to the process, these are compared with the still available vehicle stocks in distribution, production, in the daily programs, dealer settings, firm orders or forecast requirements. If a free vehicle is found in the inventory that corresponds to the buyer's wishes, the buyer order is assigned to it. Otherwise an attempt will be made to adapt the forecasts, firm orders or settlements to the buyer's request.

From the point of view of a dealer, only their own settlements (or fixed orders) are initially compared with the buyer's request. If no settlement applies, the dealer can look into the settlement stocks released for this by a neighboring dealer and try to use them to satisfy the buyer's request. This process of assigning a buyer to neighboring dealers is referred to as "locating".

The program block "settlement manipulator" carries out the entire outlined process of inventory and settlement comparison with the purchaser orders and the locating and updates the individual settlement of the dealer.

Results or issues of the settlement manipulator are updated settlements with buyer orders as well as buy orders with assigned stocks or vehicles specified in the settlements or fixed orders.

Program block "Daily program"

This program block works with the daily settlements of the dealers. Depending on the plant-specific requirements, for example the latest possible starting points for assembly, the settings are compiled into daily programs. In addition, the vehicles specified in the subscriptions are broken down into their modules. This is communicated to the suppliers as specified call quantities.

The outputs of the program block "daily program" are daily production programs for the factories as well as fixed call quantities that are communicated to the suppliers.

Program block "Production and suppliers" In this program block, the individual production facilities, the suppliers' plants and the times for assembly or delivery are reproduced in an abstract form using several model modules.

Average throughput times and throughput time fluctuations as well as the daily production capacities for the production facilities and the supplying plants are given as parameters.

In addition, the model elements have descriptions of characteristics, for example, about limit capacities, working time models, personnel numbers or other specifics of the plants, which are necessary for the forecasting and planning processes of the previously described program blocks.

The production plants simulated in rough model blocks give simplified calls to the model blocks that simulate the supply in a simplified form.

The daily programs serve as the input for this program block. The program block creates vehicles in the model that are transferred to the following program block "Distribution".

Program block "Distribution"

Just like the production facilities or the factories of the suppliers, the distribution to the dealers or directly to the buyers is abstracted in the form of model elements that simulate the mean distribution times. The extent to which the utilization of individual transport capacities must be taken into account is clarified in the conceptual phase.

Structure of an exemplary basic model

A basic model is described in greater detail below. This basic model exemplifies the start of series production and the subsequent year (for example, the year 2003) without disruptions such as strikes or supplier bottlenecks.

The model is intended to consider the US, Canada, Western Europe markets and a market that supplies the other areas. The USA and Canada markets are composed of the PPCs shown in FIGS. 26a and 26b. The markets of Western Europe and "other areas" are each represented by an importer with a 100% market share.

Equipment variant A is sold on the European market. The equipment variants B, C and D are sold on the other markets USA, Canada and "other areas".

The vehicle body includes the modeled vehicle classes and equipment.

A vehicle is described in the model described here by a family of PR numbers:

Engines, gearboxes,

Air conditioning, radio and hood.

Every vehicle also has an exterior color.

The individual PR number families are composed, for example, as indicated in FIGS. 27a and 27b.

The product tree consists of a basic vehicle description, a vehicle description "Vehicle X, US", which summarizes the common features of the equipment B, C and D, a vehicle description "Vehicle X, Europe" and the equipment variants A, B, C and D (compare figure 28).

The following constraints and prohibitions have also been defined for these vehicle descriptions:

1, 6 I engine always with manual transmission, equipment variant B always with 2.0 I engine, - equipment variant C never with 1, 4 I engine,

Equipment variant C never with 1.6 I engine, equipment variant D always with 2.0 I engine. For the years 2002 and 2003 (start of series production and the subsequent year), the sales shown in Figure 29 were forecast for vehicle X. It was also assumed that the start of vehicle X, for example, begins in KW 31 in 2002 and has the sales data specified in FIG. 30. In the model, this vehicle sales is initially divided into the two vehicle descriptions "Vehicle X, US" and "Vehicle X, Europe". These proportional vehicle sales have the course shown in FIGS. 31a and 31b.

The vehicle description "Vehicle X, Europe" or the subordinate vehicle description A is sold 100% on the European market. The vehicle description "Vehicle X, US" and the subordinate equipment variants B, C and D are reproduced with that in FIGS. 32a and 32b Distributed to the US, Canada and "other territories" markets.

The following constant distribution is also assumed for the equipment variants B, C and D:

B: 10% of the sales of the vehicle description "vehicle X, US", - C: 65% of the sales of the vehicle description "vehicle X, US",

D: 15% of the sales of the vehicle description "Vehicle X, US".

For the determination of the plant capacity, it was assumed that in plant A (outside of Europe) work from Monday to Friday in two shifts and Saturdays in one shift of 7 hours and that the days shown in Figure 33 are non-working (holidays or company holidays).

The pro rata plant output in the model was set so that the plant utilization in the model related to vehicle X averaged 85%.

The distribution was mapped as follows:

Vehicles to Europe are first transported to the overseas port (duration approximately one week). From there, shipments are made to the destination port (duration approximately three weeks). In the destination port, the vehicles finally become approximately one

Stored for a week. The distribution within Europe is not considered any further. US Distribution

Vehicles that are delivered to the "other areas" market, for example, go to Latin America (duration about three weeks).

The transports take place daily in the model. The capacity of the transports can be limited if necessary.

The following parameters were assumed for process control:

FU1 (factory assignment): 28 calendar days before ZPδ,

Daily breakdown (allocation of orders to days): 28 calendar days before ZP8, FU2 (handover to production): 14 calendar days before ZP8, FU1 and FU2 take place weekly, - forecasts are made monthly,

Volume agreements are drawn up monthly, dealer orders are placed weekly.

When forming the daily program, the probability distribution shown in FIG. 34 is assumed for the swirling of the order.

This basic model, parameterized in this way, leads (without further modification) to the results shown in FIGS. 35 to 42 for:

- Average, minimum and maximum delivery time: Figure 35, average delivery and order processing time: Figure 36, delivery time for vehicles with a specific engine: Figure 37, schedule reliability: Figure 38, weekly schedule reliability: Figure 39, - ZP8 loyalty: Figure 40,

Delivery reliability: Figure 41, inventory: Figure 42.

Various scenarios can now be developed on the basis of this basic model.

Scenario S3, for example, is characterized by an overly pessimistic sales forecast. In the basic model, the assumption is made that the forecast sales actually arrive as actual sales; in scenario S3, which assesses sales too pessimistic, this assumption is rejected. It is then examined what effects a too pessimistic sales forecast has on the process.

For this purpose, the model for this scenario S3 can be expanded, for example, in such a way that the actual sales exceed the forecast sales by a total of 20%.

Scenario S4, for example, is characterized by an overly optimistic forecast for sales.

Analogous to scenario S3, the effects that an overly optimistic forecast has on the process can be examined, for example by assuming that the forecast sales are 20% above the actual sales.

Another scenario S5 shows, for example, a strike in production.

As an example, the strike began on March 1st, 2003 and lasted a total of ten days. Two variants can be examined. In the first variant, the effects of the strike are examined without countermeasures. In a second variant, fewer orders are scheduled early because the strike was known in good time.

Further countermeasures can be taken into account in correspondingly expanded models.

In the described embodiment, the strike is mapped by a disturbance in the factory capacity.

Option one: The strike occurs without countermeasures being taken.

The simulation results of this variant are in Figures 43 to 49 for:

- average delivery and order processing time: Figure 43,

Delivery time for vehicles with a specific engine: Figure 44, Planning accuracy: Figure 45, Weekly program loyalty: Figure 46, ZP8 loyalty: Figure 47, delivery reliability: Figure 48, inventory: Figure 49

shown.

Option two: after ten days, 1,165 vehicles are delayed. In response to the strike, the target output is reduced by these 1,165 vehicles in March.

The simulation results of this second variant are shown in FIGS. 50 to 56 for:

average delivery and order processing time: Figure 50, - delivery time for vehicles with a specific engine: Figure 51,

Scheduling accuracy: Figure 52,.

Weekly program reliability: Figure 53,

ZP8 loyalty: Figure 54,

Delivery reliability: Figure 55, - Stock: Figure 56

shown.

In a further scenario S6, for example, a bottleneck in the delivery of diesel units can be examined in a simulation.

For this purpose, the model is expanded to include a supplier of this engine, for example, and the capacity is set so that the need for engines can be met exactly. The bottleneck is then mapped to a disruption in the supplier's capacity.

The embodiment of the invention is not limited to the preferred exemplary embodiments specified above. Rather, a number of variants are conceivable which make use of the arrangement and method according to the invention even in the case of fundamentally different types.

Claims

PATE N TA NSP RÜ CHE

1. Method for simulating order processing processes for the production of a complex product, in particular a motor vehicle, characterized by the following steps:

2. The method according to claim 1, characterized in that used in the automatic comparison of the demand figures in step b) according to claim 1

Records include restrictions on manufacturing locations and / or suppliers.

3. The method according to any one of claims 1 or 2, characterized in that the requirement numbers in step a) according to claim 1 are determined by

a first demand forecast is given for a first forecast period, a second demand forecast for a second forecast period is determined using stochastic methods from the first forecast and

the demand numbers are determined according to predefinable algorithms that evaluate the first and / or second demand forecast.

4. The method according to any one of claims 1 to 3, characterized in that the automatic comparison in step b) according to claim 1 comprises a correction of the demand numbers to adapt to the manufacturing and / or (manufacturing) supply capacities.

5. The method according to any one of claims 1 to 4, characterized in that the method steps a) to c) according to claim 1 comprise the following steps:

simulative generation of dealer orders for a second forecast period, preferably for three months,

- Evaluation of the preliminary demand figures and dealer orders and

Determination of an updated demand forecast for the second forecast period,

Generation of the required numbers (settlements) for the definable period, preferably a delivery week, by evaluating the approved fixed order quotas, module quotas and / or simulated at the

Dealers newly received buyer orders,

Comparison of these demand numbers (settlements) with restrictions (capacity, capacity utilization or the like) of the production site (s) and / or suppliers and assignment of the demand numbers (settlements) to the production site (s).

6. The method according to any one of claims 1 to 5, characterized in that in the automatic assignment of the demand figures to the manufacturing facilities, the demand figures of the predetermined period are divided into daily settlements.

7. The method according to any one of claims 1 to 6, characterized in that the automatic assignment of the demand figures to the manufacturing facilities comprises the compilation of daily programs for the manufacturing facilities.

8. The method according to any one of claims 6 or 7, characterized in that the automatic assignment of the required numbers to the manufacturing facilities includes the resolution of the products specified in the daily subscriptions in their modules.

9. The method according to any one of claims 1 to 8, characterized in that the required numbers include information on essential features of the products ("heavy items").

10. The method according to any one of claims 1 to 9, characterized in that the model on which the simulation is based depicts several manufacturing locations.

11. The method according to any one of claims 1 to 10, characterized in that in the model on which the simulation is based, the parameters characterizing a production site

- border capacities, - working time models and / or

- HR base

include.

12. The method according to any one of claims 1 to 11, characterized in that in the model on which the simulation is based, a differentiation is made between dealers, in particular between dealers in the domestic market and importers.

13. Procedure according to one of the. Claims 1 to 12, characterized in that in the model on which the simulation is based, distribution channels are subdivided into sub-distribution channels.

14. The method according to any one of claims 1 to 13, characterized in that the data generated by steps a) to e) according to claim 1

- quantitative assessments of process designs,

Assessment of strategies such as fault management, freezing times of orders, delivery times, delivery reliability, - utilization of means of transport and / or

costs

include.

15. The method according to any one of claims 1 to 14, characterized in that data from databases of real systems, in particular from databases of dealers and / or manufacturing plants, are automatically evaluated during the method.

16. Simulation system which comprises the modules "forecast", "fixed orders", settlement "," production "and" distribution ", the modules, controlled by a computer program implemented on a computer system, interacting in such a way that the following steps can be carried out :

d) execution of a simulation of production and / or delivery for production based on the assignment made in step c), e) automatic determination of the distribution channels and execution of a simulation of the distribution (s) of finished products from the plants to the delivery locations,

17. Simulation system according to claim 16, characterized in that the simulation system has interfaces to databases of real systems, such as databases of dealers and / or manufacturing facilities.

18. Computer program product, which comprises a computer-readable storage medium on which a program is stored which enables a computer after it has been loaded into the memory of the computer, a method for simulating order processing processes for producing a complex product, in particular one Motor vehicle to perform, the simulation comprising the method steps according to any one of claims 1 to 15.

19. Computer-readable storage medium on which a program is stored, which, after it has been loaded into the memory of the computer, enables a computer to carry out a method for simulating order processing processes for producing a complex product, in particular a motor vehicle, the simulation comprises the method steps according to one of claims 1 to 15.

20. Use of a method for simulating order processing processes according to one of claims 1 to 15 or a simulation system according to one of claims 16 or 17 for determining planning data such as optimization potential, decision alternatives, key figures for delivery time or delivery reliability, means of transport utilization, costs or the like.

21. Planning data such as optimization potentials, decision alternatives, key figures for delivery time or delivery reliability, means of transport utilization, costs or the like, which are generated by a method for simulating order processing processes according to one of claims 1 to 15 or by a simulation system according to one of the

Claims 16 or 17 have been made available.