TW200300536A - Sales forecasting apparatus and sales forecasting method - Google Patents

Abstract

A sales forecasting computer 10 obtains a sales result of respective goods to be handled from a database 8, sets respective goods into groups by using a sales period of respective goods as at least one index, based on the sales result, decides to which group the goods to be forecast belong, based on a sales period of the goods to be forecast, and calculates a sales forecast of the goods to be handled, based on a sales result of the group of goods handled during a past period corresponding to the forecast period having a similarity with the period to be forecast.

Description

200300536 V. Description of the invention (1) 1. [Technical field to which the invention belongs] The present invention is a system for predicting goods and services in a warehouse. 2. [Previous Technology] In order to avoid too many or insufficient products for efficient & selling, it is to predict the sales amount to determine the order quantity. x? Lock sales vendor For example, in Japanese Patent Application Laid-Open No. 5-1 2 0 3 1 4 _ The following sales forecast system. On this system, ^ = 上 ’揭 = has past sales to predict future sales. Based on the monthly sales of the target product, sales of the most recent three months in the country, and Qiankou. In the same month (3 months), the sales consultants are all considered to predict the salespersons', which can improve the accuracy of the sales forecast. However, on the premise of this system, the premise of the previous car, the title, and the beneficial method for the goods that have not been sold, is to make a two-item, one-to-one prediction. = \ This pure is not applicable to the industry where the product being sold is subject to change or frequently invests in new products. This is the problem. This: Each product refers to its own past sales results one by one. Two: Therefore, when the number of predicted target products is large, there is a problem that it is difficult to perform sales forecasting with :; In the Japanese Patent Application Laid-Open Publication No. 8-2 78959, the support system has disclosed a system for forecasting sales of products that have a beneficial effect: the best product. In this system, attributes such as :,, mouth, pattern, size, grade and so on are classified in detail and recorded. For products without actual sales, select the products with the most similar attributes from the records that have been recorded, and make predictions based on the sales performance of the products.

$ 6 pages 200300536 V. Description of the invention (2) Measured sales. If there is a difference in performance, the forecast will be used. But the detailed properties, the problem. 3. [In the invention, the actual sales performance of the present invention is non-tradable, and the sales forecast is based on the sales performance of the present invention, which is at least as a group of transaction objects, and the sales period for the present invention is calculated based on the corresponding delivery period. The test device can reach the predicted transaction pair group. From the above forecast period and j, the sales forecast is on time. Then, the forecast is revised. … 仃 'When forecasting and realizing the system', it is necessary to record the sales volume of non-traffic 纟 soil in order to improve the accuracy of the forecast. 铮 M γ * and the question of the huge number of attributes that record the record storage of goods] Purpose In solving the above problems and avoiding the complication of processing, and St: a kind of free sales forecasting device. b. Enter an appropriate forecast for the sales amount of locks and locks. Based on the above-mentioned sales performance, the index of the second exchange object 'plants each transaction object group, and ^ the sales period at the time of sales to determine the forecast of the transaction target. The group of transactions that the transaction object belongs to and the group #period and the similar sales that are similar to each other: the group's lock-up sales performance is based on the characteristics of The object records the grouping of each group, and the sales amount is ":: = broken; ... as if it belongs to-^ has a transaction object with a similar Γ pair, and the transaction object in the past corresponding period 200300536 The relative sales ratio is calculated, and the future sales can be predicted. It is characterized by the use of the previous period. Therefore, it can accurately reflect; ^

On page 8, we use the sales performance to calculate the sales that can be oversold based on the sales units predicted by the sales unit in a single period of time. The amount is increased, and each of them is measured. Because of this, it sets the process of sales, its progress and progress to determine which group of each transaction target group, according to the time of sale according to the time of the elephant, there is no elephant's five, send friends description (3 ) Group sales forecast, that is, the actual results of the forecast are grouped into groups. In addition, the invention is added to make the sales forecast for each given plural unit. The present invention seldom calculates the ratio of the ratio in the time period of the present invention and the forecast of the menstrual period. Therefore, it is considered that during the period of passing the unit of the present invention, the target of the sales period is the amount forecast. The sales volume forecast period is complicated by the clear classification of the number of transaction objects. During the sales pre-sales volume forecast period, the relative sales volume is changed. The sales volume is forecasted. According to the actual sales forecast in the time series, when the unit characteristics are set as the unit to be determined, the transaction can be turned into a group, and the sales of the group to the Hungarian group will over-predict the accuracy of the transaction object. The period $ t 'is calculated based on the time-series land sales feature as the sales forecast is the ratio of sales to sales. So at least the tendency. Characteristic is based on the relative sales of plural units per unit period. Sales advance forecast without sales. Unit sales without past period. 200300536 --- V. Description of the invention (4) Changes in sales performance trends The sales sales forecast of the present invention. The period is used as a passing and forecasting device, which is characterized by the use of plural units and thus generates sales per period. Therefore, even if there is a sudden change in the sales forecast due to a sudden special change, the sales can be reflected without being affected by this change i 为 ί = i is a plural unit ㈣, so the lock-up sales amount is preliminarily changed. In order to eliminate special factors, the sales of the invention according to the present invention, the sales volume of each period, and the forecasting device are characterized by the passing of the relevant unit, and when the purchase is judged based on the purchase and the lock-up, it is estimated that the inventory is not correct. Correct the sales amount for every # 犄, and use the above sales performance to modify: Therefore, == to predict the sales amount.亍 时期 二 二 ； ； ； 产生 产生 产生 产生 隹 隹 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 产生 中 中 中 中 中 中 中 中 中 中 中 中 的 中 中 中 中 的 的 的 的 中 的 的 的 的 的 变 in the reaction period of the actual sales of lock-up sales; :: π 隹 is characterized by being used in the plural for accurate amount prediction. Therefore, “& ^ Lu Ming ’s sales forecasting device can be used to advance sales of Jike's sales in the period before or after or before and after Jr. soil: Sales: Considering this, we will also consider investing in cats, Heading from the sales forecast of shellfish, shellfish. Due to the time before and after, the %% predicted by the parent of the parent-easy object belongs to the group of the gangsters, or the period of fools. Buying period: The performance is used to make sales forecasts. Also: Month: The stable sales forecast that is affected by the sales of other groups.

Page 9 The sales forecast of the present invention is set at 200300536 during the sales forecast period. This transaction is a sales forecast. Testing equipment, easy-to-sell sales, testing equipment, ordering. Completed billing, necessary purchases, testing, and installation. Time: can be five, the description of the invention (5) the entire soil of the complex period of the complex period of the sales object group corresponding to the calculation of each unit period. Therefore, based on the overall time-locked sales amount Sales to staff sales forecast to calculate the order information of the purchase. Therefore, according to the predicted sales amount of the present invention, it is preset to be shorter than between the transaction objects. Therefore, considering the actual sales results, it can be determined that the sales forecast of the present invention is considered to be different from the above-mentioned purchase amount of the biased branch. Based on the forecast of future scale sales, the amount of deviation is considered. The change in sales performance with the transaction object to which the transaction object belongs, ^ sales forecast.疋 ’can carry out the purchase amount for each unit period and # inches according to the calculated lock requirement, and then output the monthly purchase amount to indicate the necessary purchase amount. / 、 Special attack is required to change the unit period from a predetermined unit to [] the required departure period from the parent product = the purchase schedule and sales of the target object. T 2 attack is required for the single standing period of the calculation unit of the required purchase amount. The necessary purchase quantity, the above accurately determines the characteristics of the necessary purchase A. Bit period: · The predecessor system in the unit period determined by the forecast of a single predetermined number of future stocks, and the sales of the invention according to the forecast library The amount of pre-purchase, the necessary purchase amount of the period, is more measured than the unit period, and the total amount of sales is equal or strict to determine the above-mentioned purchase measurement device.

Page 10 V. Description of the invention (6) Therefore, the necessary purchase amount is calculated based on sales, ^ ~. ^, Prediction to determine the appropriate degree, by adjusting B Youbu 'according to the prediction,' M ΑΑ can calculate the period of prediction. ; ^ Has an appropriate purchase volume than the accuracy of the I :::? Degree and prediction. (Or the established relationship) the inventory amount is as follows: the sales box of the present invention has sales 隹 *, forecasting devices, and ϋ * to achieve accurate results. : Changes. Therefore, it can be classified according to the sales of the month = the month of the month, according to the sales / attributes of the present invention in the grouping period. When the object can be sold with high accuracy based on the combination of the 4 anti-Bayesian S-periods and the group's lock-up sales amount, when it can be sold with high accuracy: :: sales forecast device, ... : Grouping from the sales period from the beginning of the sales period and the prediction of the amount box and the base of the invention. Γ Achievements, roots: It is characterized by obtaining the transaction objects of each transaction object, whether the board data object corresponds Any similarity past time, the forecast period of the forecast transaction target, and the sales performance of the transaction target sales of the similar transaction target are used to calculate the forecast. That is, the root copy Shanghai Cooperation ^ Predicted trading pair = 1 buying period to classify each trading object, determine whether the process corresponds to any existing trading object, according to

200300536 V. Description of the invention (Ό The trading target ’s date of sale is used to classify the transaction's prediction of accuracy. Due to the excessive increase in classification, and in this book, the invention of the method can also allow the so-called In terms of products and services. Targets. The so-called "sale of the past and the future. The amount of forecast is not equal to the amount of demand. The absolute number also includes the so-called" sale of the so-called "root. , Including the so-called "record RAM, floppy disk, CD_ at the end of the sale. In addition, it also contains the actual results of electricity sales, so in addition, there is a document that can be used to determine the function. The easy sales target is only the commercial price The concept is the same as the 'ROM > words loop of a program such as the sales volume based on the actual sales data, the next year, etc.

^ Perform sales forecasts. Since the classification of this transaction object is clear at the time of sales, even if the number of processing objects increases, there is no reason that the transaction may become complicated. =? The invention of the method can be mastered by the invention of the method :: 垂 ^ 明. In addition to the dagger, the above-mentioned invention ’s current program is recorded in a computer record ^ is deducted from the business that predicts sales in the present invention. Clothing products are equivalent to the transaction test "refers to predicting unknown sales, regardless of The sales forecast of products also includes the concept of the use of electrical appliances, and the concept of external sales is not only the same as the index expressed on the method (the relative ratio is expressed in the same way). "+ 疋 refers to the quantity actually sold. 4, "Grouping" refers to the concept of grouping the parent-exchange objects according to the transaction object, not only the start time of the sale and the week of the sale, but also the concept of grouping. ",,, and recorded media" refers to recording media such as RQM, memory cards, and hard disks. Concept of communication media such as transportation and transportation. No 200300536 — ~ — ^. V. Description of the invention (8) 疋 像 人 c Ρ ϋ Connects directly to execute the recorded program, and it includes a hard-loaded δ record that will be executed once installed on the hard disk, etc. Down CD-the concept of a record media that is waiting. "Taxi taxis, household programs" are not only programs that can be directly executed from the CPU, but also the source programs, compressed programs, encrypted programs, etc .. [Implementation] 1 · The overall picture of the system 1 The system configuration is shown in the sales forecasting system according to one embodiment of the present invention when it is first applied to the management of clothing products. In this implementation method, a sales forecasting device 10 is constructed based on the sales forecasting computer 10. The host receives the purchase information and materials from the logistics that centrally enters and leaves the warehouse and keeps the goods. In addition, in stores nationwide: = Cash register device 4 ′ transmits the sales data to the host computer 6. In addition, the computer 3 of the manufacturer makes an order. In this implementation method, although the data is transferred to the host computer 6, the data can also be transferred via a recording medium such as a floppy disk. In addition, data printed on paper or the like is transferred to the host computer 6 in turn. Furthermore, the data ′ output from the host computer 6 can also be output using a recording medium or used in rotation. · 笊 The host 6 accepts these data, performs centralized calculation of data at a fixed time every day, and communicates with the database server to update the database 8. The sales forecast computer 10 is connected to the database feeder 9 through the UN,

Page 13 200300536 V. Description of the invention (9) Then according to accumulation +-owe, + reason. Animals and materials in the shellfish storehouse 8 are used to make sales predictions. 2 The system outline 31 is ^ The system shown in Figure 1 is schematically shown for the purpose of sales forecast !: Chromium: Γ The product flow and time The evolution of information processes. ①: Department 18 2: 隹: Γ :: Sales are real. This section 18 is to determine the forecast of each product: the target) depends on the past sales σ, ^ of the group to which the product belongs to make sales forecasts. ② Headquarter 18 places the first order with manufacturer 19 based on the calculated sales forecast. ③ The manufacturer 19 produces the goods according to the eight ordering instructions. It takes 12 ° to the logistics center. The less-wide logistics center 12 transmits the incoming and outgoing materials to and from the logistics center 12 from the logistics center 12 according to the instructions of the headquarters 18. ⑤ At Kushiro Station, 1/1 1 η, taxis will be sold; £ 1 4 1 R + ° will be used to purchase the goods. Youkou stores 1 and 4 will send the sales to the headquarters 丨 8 ⑥ On headquarters 18, according to the sales start period, the merchants will record the changes in the sales of each group; 夂 sales data (the past sales performance ) And each quotient = 18 to predict future product sales based on the change in sales volume of σ quotient, and to predict future product sales. The genus group: based on the sales forecast calculated in the past, based on the past, the manufacturer and the department are repeated based on the results. Perform Θ ~ ⑦. . ° Place additional order. As described above, it is possible to perform high accuracy 纟 + as described below, ‘can reduce the library 1 test’ according to the above, reduce the generation of feathers and reduce sales 200300536 V. Invention Description (ο) Lost opportunity. 3. Host machine 3〇 ^ CD-ROM „324; ΓΪΓΙ8. ^ The second system is to communicate with the logistics center computer 2, the store cash register device, 99 the manufacturer's computer 3, the sales forecast electricity, and the communication object. .Hardness 2 seems to record more operating systems and data sets: outside, = there is a construction database 8 in the ribs. In the database 8 = yes =;: the sales of Zhongshibi. Inventory, data and products The program for the calculation and processing of the main Q data of commodities such as prices is installed into the hard disk drive 22 via the CD-drive 隹 Φ,: _ 38, etc. Figure 4 is a flowchart showing the program for the data centralized calculation and processing. ^ S1 上 'CPU2G receives the sales data transmitted from the store cash register of each store # (Which product sold how much \ Equipment / Electricity: 2 the inbound and outbound data (which one

The next order (the number of orders for which product is placed) Jin: J: Poor: Stored in the database 8 of the database feeder 9. This: place = reverse reversal until reaching a predetermined time (step%). As soon as the given time is reached,% is counted (Qing). -Generally, it is a centralized calculation of the data received. There are additional sales data and in and out 200300536 V. Description of the invention (11) ^ The night of the warehouse is used to perform centralized calculation of this data. In the data collection calculation process, 'the sales data of each product in each store is collected and calculated differently', and then the calculated data is recorded in the hard disk drive 22 as the current day's sales. Similarly, the inbound data and outbound data of each product are calculated in the database 8 and stored in the database 8 of the database server 9. As a result, daily sales, inbound and outbound data is stored. It also calculates the total weekly sales (total sales from the beginning of the week) of each product, the total number of purchases so far, the total number of sales $, the number of stocks, and so on. All these data are recorded in the database 8 of the database server 9 (step $ 4). Fig. 5 shows an example of recorded data. As shown in the figure, each product is planned according to a combination of a variety and a product name number. In this implementation method, there is a case where a plurality of products are included in a variety, and a description will be given below when a plurality of varieties exist. FIG. 5 shows an example of stored sales and inventory data. The number of locks for each week is ^ per item from Monday to Sunday of the week. Miles are different. In the figure, the product with the product number "6 2 4 2 1" indicates "0 photos" in the first week, ... "30 photos" in the 30th week, "40 photos" in the 31st week, and "48 photos" in the 32nd week " The cumulative number of sales for the week (here, week 32) is determined on the evening of the week. f. The purchase quantity is the cumulative purchase quantity of the product. The cumulative sales volume is the cumulative sales volume to date. The inventory quantity is calculated based on the cumulative β ^% quantity-sales quantity-an adjusted quantity. The number of bins is updated every day. In addition, the so-called adjusted quantity

Page 16 200300536 V. Description of the invention (12) ΐ: In terms of the amount of stock reduction, such as damaged goods, leftovers, etc. The number of mouth temples. Such merchandise information is also transmitted from various retailers. η: η, which is the week in which the initial sales volume is not 0 (that is, the week when the brother bought it once). In the 8th storehouse, not only records the sales this year as described above. Figure 6 shows that $ 去 贝 # has been recorded before the last year. See the same results as examples. Also, although not shown, Yemen ’s sales performance data before ‘,,, and Ji ’s are higher than last year. 4 · Computer hardware structure for predicting sales The main component of the disk is the computer hardware structure. Basic structure: beam mu: In law, at the beginning of the year (or in the previous year, the min.), The weekly index for predicting sales is calculated, and it is recorded on the hard disk drive 52. in. U number (Steel sales index weekly index calculation process is shown in Figure 8 shows the weekly index calculation is performed at the beginning of the year. Baixian, CPU50 access to the database 8 from last year's household " spoon machine private sales number More product information (step j again. Performance data 'arrange the products with a large number of sales in order. People implement this method with a cumulative composition of 70% or more of the products. So, select the products with a large number of sales. Sales performance data, = pull out and hold j M

Page 17 200300536 V. Issued clearly (13) Changes in the sales performance of the goods by category and sales. In Figure 9A, the table is an example of actual sales data for the previous year. Next, this sales performance capital will be grouped at the beginning of the week (step S12 ',). Here, the term "item" refers to a product group such as skirts, jackets, and pants. For example, in the data of Fig. 9A, although only the skirt data of the item number "〇1" is shown, the same information exists for other items such as jackets and pants. In addition, the so-called sales start week refers to the k and a of which the product starts to sell to women. That is, for the week when one or more copies are sold. In the data of FIG. 9A, it is common that the product ^ sales start week of the first line and the second line is "01" (that is, the first week). Therefore, the products in the first and second rows are considered as one group, and their sales performance is calculated together. The resources obtained by grouping like this are shown in Figure 9B. In this way, last year's results of grouping of item 7 sales start week can be obtained. After that, cpU50 uses the group anchor # of the week before and after as the sales start week. Contributing-Calculate the Ping Shao of the actual performance data of each week and sell it. That is, use the following formula

Savg (Item, ws, Wf) (step S1 to calculate the average of sales performance .;

Savg (Item, Ws, W f)-(sn tpm w…, ΐ e m, W s-1 w f) + S (11 e m W s W Ή + S Γ

Item, Ws + 1, Wf)) / 3 u 卞… τem, ws, wr) + w Here ‘I tem represents the item number of the item. The group is defined by It tem and Ws 3 1. Bang, and Wf 疋 will start the week of sale as the "week 1" ^ week number osav (Ite w w 十 、 联 —cognizance is selected based on Item, Ws

Page 18 200300536 V. Description of the invention (14) The average of sales performance on the Wf week in the group. s (x, Y, z). The ironing sales performance of the sales start week in the second week of the gamma group.疋 口 口 Figure 10 table does not include the average sales performance for the week before and after. In addition, although the above indicates only the data up to the 7th week and the 6th, when the actual sales performance exists, such data is also recorded. It ’s so flat; ’Even if it ’s sold out because of unexpected reasons, it ’s another: In the method of application, as described later, the sales of goods are forecasted based on the seller = Zhou Weitu 4. Therefore, it is best that all weeks are the sales start week of any of your product groups. However, for items that cannot be sold as actual performance data, the week of the specified week is the data of the start of the week. For example, as shown in Figs. G A and B, there are products in which the fourth week is the start week. However, by calculating the average sales performance of the group including the 4 weeks as the start week, data as shown in FIG. 10A can be obtained. Again, all week

The total of Savg (Item, Ws, Wf) is recorded as T (Item, Ws). ^ As described above, after calculating the average sales performance as shown in FIG. 10A, the weekly index (sales index) for each group is calculated (step S14). The calculation of index (item, ws, wf) in Zhou Caiyue number is performed by the following calculation formula. Cut Inde ^ Item, Ws, W * f) -Savg (Item, Ws, Wf) / T (Item, Ws) FIG. 1013 shows an example of the index (Item, Ws, Wf) calculated as described above. In addition, the 'CPU50' is the same as the above, together with the weekly index of the previous year, the average weekly index of 2 years ago and the weekly index of 3 years ago (step

Page 19 200300536 V. Description of the Invention (15) 15). In this implementation method, the weekly index of the previous year, 2 years ago, and 3 years ago are weighted by 5:32 to calculate the average value. Figure 1 丨 shows the weekly index after an average of three years of injury. This weighted average weekly index is recorded in HDD 52. " & The weekly index with an increase of 3 years on average, in addition to excluding ^ = sales changes due to special requirements, can also be used to reflect the recent sales tendency. As described above, in the hard disk drive 52 of the computer 10 for predicting sales, the average weekly index for increasing sales is recorded. Order quantity calculation processing 6 · Sales forecast 6 · 1 condition setting pin 1 2 indicates the jaw forecasting of the jaw sales recorded on the hard disk drive 52 of the computer 10 for forecasting sales • The flow of the order quantity calculation processing program Illustration. Immediately before I *, the CPU 50 displays the setting of the conditions on the screen 54 and reminds (step). FIG. 13 shows a day surface for condition setting. As a helmet on this day, you can use the preparation period (the period from the time the order is placed to the time required to enter the warehouse) as the week number to enter the material of the product. The main materials used are: Commercial: Weaving "materials of materials" "C" is mainly used after cutting products; 眢 ...: ▲ J "is an example of using materials such as sportswear 2" table Not shoes, scarves, hats, etc. Enter the preparation period of 4 weeks on the picture "「 Ρ 」, and enter 3 weeks in the" relevant "into the gate of Rumen-Shi Γ worship% u _ μ / i In 3 relevant rounds in the" j " "Enter 3 weeks of preparation for" groceries "

Page 20 200300536 V. Invention. Explanation (16) / month 1 These preparation periods are used when deciding the order quantity. In addition, you can choose "effective" "no 1 ^" in the "defective goods correction". Here, the so-called "defective goods compensation" refers to the calculation of the sales volume of the previous week in order to predict sales. In the case that the storefront cannot be sold because of defective products ", the sales performance is corrected. If it is set to "Enable", the ππ correction will be performed.设定 If it is set to "Invalid", the defective products will not be corrected. Eight A and ‘the“ forecast inventory week number ”column indicates that there are quite a few weeks. In Fig. I, we specify the amount of stock that corresponds to the sales volume of 4 weeks. Each of the above-mentioned poor news is recorded in the hard disk drive 52, and is used when the forecast sales process determines the order quantity. 6 · Predicted sales amount (step islf) After the condition setting as described above is performed, the sales amount is predicted. Figures 14 and 15 show the flowchart of the sales forecast process. In this implementation party: Separate the sales forecast of the product in the state of no sales performance ($ 沾 U sales forecast) from the sales performance of the product obtained: the lock-up sales forecast (additional sales forecast) deal with. Bei Zhi's 1st sales forecast, CPU, a. The tested products, CPU50 from ^ Baixian, related to the pre-sales of this sales (Niu Su 3 Λ Λ to obtain the advance forecast for this year's staff) (Vstep 31) λ sales forecast is based on the sales target of this product

200300536 V. Description of the invention (17) ^ · 'The target and product characteristics are determined on the basis of the tomb and recorded in the database 8. Next, the sales of the goods obtained from the database 8 are scheduled to begin. The sales start date is also determined and recorded based on the characteristics of the product. According to the sales start scheduled date and the item classification of the product ^ $ to determine which group this product belongs to, obtain the weighted average weekly index of the group from the hard disk drive 5 2 (step S 3 2). After that, the CPU 50 predicts the lock-up S p (g 0 d, w f) for each week based on the sales schedule of the product and the average week number (step s 3 3). ° SPCgood, Wf) = ST X Index (11em, Ws, Wf). Here, sp (g00d, Wf) is the sales forecast number 4 for the product g00d from the scheduled week to the Wf week. , ST- is the flfl ^ lndexdte ,, Ws, Wf of the merchant's nod), ^ t ^ ^ I t: is the item category to which the merchant sg00d belongs, the sales start, and the sales of the product g00d start the scheduled week. That is, according to the weighted flat, the full predetermined sales are allocated to each week and the sales per week are generated. The CPU 50 records this sales forecast in the form of hard print and data to output: (step S34). The processing of: is executed for the target of sales forecast. Therefore, “阏”: After the predetermined sales amount is determined, it is recorded in the hard disk drive 52: = no for each item's sales forecast code. In the picture in Qian 2, the SKU is the code for each specific product, which is calculated in this way. 200300536 V. Description of the invention (18) _ When the sales performance is generated, the sales are different. Figure 16 shows the output of this processing. Based on the first day of each week, it is recalculated on the first day of each week (in every two days. Figure 16 sales forecast processing usually CPU 50 first from the database 8 to 2 Period —) is executed. S41). In addition, according to the actual performance data of the obtained T α quotient ... (the steps indicate the obtained actual performance data—for example. This is different from the digestibility rate, etc.) Stores and items are recorded in the database 8. After that, the computer server calculates them in a centralized manner, and records the factory product number "" s ζ specifications and colors, according to these can special C = L "points ^ indicates the product number, product number, The specifications and colors become the top 4 solid products. The combination of the above records the ".", "C," and "11." The standard selling price at the storefront of "material name." Beginning week. "Previous generation" is In the column of "5 weeks ago", record 5 weeks of worship 4 weeks ago ", 3 weeks ago," the number of sales in the week of ugly worship. Last, 3 weeks ago, last week, last week, work week " "Accumulated for 4 weeks" is the number of sales of the product until the 2nd week = "Acquisition of" Sales for purchase "is the accumulation of the number of items to the ::: purchase. Total." Weekend inventory "is the previous weekend Hours: Two: The number of sales sheets for two is exhausted, and I is scheduled to be in the cargo dagger. Purchase "Zhou purchase" "3 Dai Zhou push.", Chen number 1. Similarly, "lower

J factory j will be purchased after 4 weeks "are the number of purchases scheduled for next week and 4 weeks, respectively" CPU50 is based on the cumulative purchase and second count, based on the purchase Zhang 200300536 V. Description of invention (19)- Within the number, figure out what percentage was sold (digestion rate), and then record it on the "Cumulative Elimination Rate" block. Furthermore, based on the number of purchases last week and the number of sales last week, based on the number of purchases, calculate how much < less is sold (single weekly elimination rate), and then record in " Single ^ sales rate ". In this implementation method, the weekend inventory is added to last week's sales and then last week's sales are divided to obtain the weekly digestibility. < Next, the CPU 50 determines that the defective correction set in Fig. 13 is valid (step S42). If it is invalid, the defective product will not be corrected, I step, 45. When the defective product correction is set to be effective, the defective product correction is performed as follows. T ^ In any of the 2nd stores, the product will be sold 'cannot be sold due to insufficient inventory (defective goods :: quilt condition, use this performance: 隹 because it can not reflect the original misstatement forecast . Therefore, here. With this actual sales performance, the wrong result. Under the actual conditions, Xinzheng used the sales forecast CPU50 to presumably call it the defective product correction. In this implementation method: 2 Whether there is production, the existence of such underdetermined inferior goods. That is, if the conversion rate and the digestion rate per week are more than 60%: Second, "the conversion rate is more than 60%. Defective goods are generated in the store. In the case of defective goods, step S45 is performed. Judgment (step S44). The following formula is used to correct the sales performance of last week. Last week sales ^ (last week sales / 〇6)-f; (Weekend inventory + 200300536

V. Description of the invention (20) Salesperson of the previous week)) X 〇 6 + Sales of the previous week Therefore, the sales of the goods in the first line of Figure 17 A were revised to 3 8 6 last week column). In addition, last week's sales of products that did not meet the above conditions 4, 2, and 3 were not corrected, 277, 241.丨) Sales 疋 Next, the CPU 50 calculates n sales prediction value SG (good, η) for each product with the following formula. SG (go〇d, n)-(SR (g〇〇d) / Index (Item, Ws, Wc-1)) γ

Inded (Item, Ws, Wc + n) λ Here, SR (good) is the sales performance of the previous week (after repairing defective products, it is the actual performance after correction). Inde} i (Item, Ws, Wc—}) is the weekly index of commodity labor Item, Wc-1 of the week before the start of week group Ws. Representation,

Wc + n represents the week after n weeks from this week. In this implementation method, this ° and / are the 39th week. In addition, η was changed from 0 to 6, and sales forecasts for each week from this week's envy were made. Perform this sales staffing test for each product. Fig. 17B shows an example of a weekly index (three-year weighted average) used for prediction. In addition, FIG. 18 shows the calculated sales forecast for each week. Calculate the sales forecast from "39 weeks" (this week) to "45 weeks". & 6.3 Calculating the necessary order quantity As described above, the CPU 50 calculates the necessary order quantity in the initial and additional weekly sales forecast (step S23 in Fig. 12). , "

200300536 ♦, .. Ά. V. Description of the invention (21)… = idea. Items placed this week (39 weeks) =. 2 weeks (controllable chest week). This preparation time is based on the current implementation method, starting from this week (from 3 to 9 weeks) a week before the product preparation time LT ⑷ ❹ mouth: Haishang's forecasted sales consistent inventory to determine this week's order 'Quantity': 4 weeks — Inventory with consistent sales forecast for several weeks, we = No, then "4". In the case of Λ Λυ // · 11 ', the expected inventory week number is taken as the inventory amount of sales up to that point.隹 仟 ”to compensate for the 4 weeks. ^ · F = indicates the flow chart of the calculation procedure for the order quantity. CPUs η ;; ten, t weeks (39 weeks) to control the first week (42 week sales forecast, and obtained "773, scoop quotient: δ ten 39 weeks ~ 41, followed by (7 3") It is recorded. The predetermined number (step 2 is regarded as "Processing 1". Also., The number of the second root market completed early but not yet purchased, single record and purchase record are available; recorded in the data bank, The “work 1” of each product is “1100” to be out of the box \ predetermined number. In FIG. 18, “plus” is added. # Pre-restocked quantity and record the stored number as: Inventory 1? At the beginning of the week (42 weeks), the library "Last Weekend Inventory" plus "Exit Plus (for work)". "Inventory i" is available

Calculate "Processing 1" after subtracting "Sales Forecast i" 200300536 V. Description of Invention (22). Next, from the sales forecast for the controllable period (step ^ week), the sales forecast for the inventory week is aggregated. Missing ^^. That is, a total of 42 weeks to 45 inventory totals "refers to the current relevant # = 秘 5). Here, in terms of the number of "processed states." Also ^ σ ϋ "Processing of Π (completed order is the processing scheduled to be received after the first week of purchase., Process 2") means that if "Inventory 1" is greater than 〇 的 红 ”一 (" Inventory 1 "+" Add "Order Box" (Steps S56, S5;), cm〇 Calculate the Order Box with the following formula "-" Sales Forecast 2 Job Z ") In addition, If "inventory 丨", in the "single frame" (steps S56, S58) y ', use the following formula to calculate "the above calculation result i obtained: sale y: then 2"-"processing 2" as the number of orders If the "early morning frame" is greater than 1, then this time, Fig. 18 shows the calculated = frame "(step S59, Yang). The quotient of one line. The mouth of the mouth, the mouth of the second line: the morning frame" and the number of orders -Example. No. 270 orders should be placed, but no orders were placed. The order amount after the product calculation on the third line is also printed. In addition, Xu Youbei took turns to store and record. The manufacturer accepts this transmission = host 6 to transmit to the manufacturer 3. The production is also on the above? "" Production (refer to the figure ". ' Τ " ..., Λ-V. Description of the Invention (23) Early Box "as the upper limit, the order will be determined by the person. In addition, on the listing, the second line is shown online, but the ordering finger can also be used to , j Clothing manufacturer 3 sends the order finger%. The system and printing papers are sent 7 · It (1: Ming, product, is a heavy single tube (2: the amount, the yield (3) sales (4) sales Method to order his implementation method) In the above implementation method, although from other production ㈣ = commodity is, for example, industrial products, processing houses, for example: quotient σ of effective management quotient quot For example, it is also applicable to products placed in convenience stores. In the above implementation method, the order is calculated based on the sales forecast: Yes, the production can be determined based on the sales forecast. 4. The purchase amount can be used in the above implementation method. , # The sales forecast number, also ®, the sales amount, etc. ® In the above implementation method, even if the initial order for the product is calculated based on the forecast, the initial order amount can also be other To determine (for example, the overall decomposition of the day, etc.), only the additional amount Determined by sales forecast.

200300536 Brief description of the drawings Figure 1: The overall view of the system. Figure 2: A schematic diagram showing a business model. Figure 3: A diagram showing the hardware configuration of the host. Figure 4 · Flow chart showing the processing of lean collection. Figure 5: A graph showing sales and inventory data. Figure 6: Sales and inventory data. Figure 7: A hardware configuration diagram of a sales forecast computer. Figure 8 · Shows a flowchart of the process of creating a lock-sell index.

Figure 9: A graph showing actual sales data. Fig. 10: A diagram showing a course for calculating a weekly index. Figure 11: An example of a weekly index. : Fig. 12: Flow chart showing sales forecast and order quantity calculation processing. Figure 13: A diagram showing the content of condition setting. Figure 14: A flowchart showing the first sales forecast process. Figure 15: A forecast chart showing sales by product. Fig. 16 shows a flow chart of sales forecast processing at the time of addition. Figure 17: A graph showing poor performance and weekly index. Figure 18: A graph showing sales forecast and order information.

Fig. 19 is a diagram schematically showing a calculation method of the order quantity. Fig. 20: A flowchart showing the calculation of the required order quantity. Component symbol description: 2 ~ logistics center computer 3 ~ manufacturer

Page 29 200300536 Schematic Description

4 ~ Shop cash register device 6 ~ Host 8 ~ Database 9 ~ Database server 1 0 ~ Sales forecasting computer 1 2 ~ Logistics center 1 4 ~ Shop 1 6 ~ Shop 1 8 ~ Headquarters

1 9 ~ Manufacturer 20 ~ CPU 2 2 ~ Hard disk 24 ~ Display 2 6 ~ Communication section 28 ~ Memory 30 ~ Keyboard / Mouse 32 ~ CD-ROM drive 3 4 ~ Listing machine ·

3 6 ~ F D drive 50 ~ CPU 5 2 ~ Hard disk 54 ~ Display 5 6 ~ Communication section 58 ~ Memory

Page 30 200300536

Page 31

Claims (1)

ZUUJUUDJO VI. Application scope of patents1. A sales forecast The sales of Jun Department, x, based on the past sales performance to predict the future characteristics are: Obtaining the transaction results of each transaction object according to the above-mentioned sales actual I; Grouping of small and alpha parent objects to use as the delivery index for the prediction of each delivery basis; The measured transaction object belongs to ‘that — a group = the sales period, to judge the predecessor! For the prediction of delivery, The correspondence of similar pasts is calculated by "having a broad basis with the period when the forecast was made; the sales of the easy target group are continued 2. A sales forecast device predicts the sales value of the wall. Sales:, x Past sales performance to predict the future Its characteristics are: Pairs of indicators: i ::: Periods are grouped, at least as a record department; ', access to each group of records The actual sales of a person depend on which group the transaction pair is predicted to belong to; and, t /, to determine the transaction to i 隹 ^ ,,; the transaction target for which the prediction is made, from the i / R Ji you stand :;: Γ Amount of actual results, lock transactions * ^? Sales forecast value. , Beibei 勹 & to calculate the sales forecasting device _____ ', 肀 of the 5 Θ patent circle, and perform J: Page 32, 200,300,536 V: VI. Forecast of the date of application for the patent scope & the predetermined period of each commission, the unit period of multiple units had 4 periods, at each of the predetermined units 4. If the scope of patent application is the second ^ Prediction of platinum f. The forecast period is each predetermined sales amount forecasting device, in which the previous period is carried out, the unit period is the plural period, and the period is at each predetermined unit. prediction. The sales amount prediction is to calculate at least the relative sales ratio in each of the single imitation: amount prediction devices', the above-mentioned sales. The sales forecast is calculated at least in each unit forecasting device, where ‘the above-mentioned sales 7. The relative sales ratio of the sales order of item 5 in the scope of patent application. In the plural unit period described above ::: measuring device 'wherein, according to the above, the sales ratio and the relative sales ratio of the unit period without the unit period after the actual performance period, the sales ratio and the sales forecast Amount forecast. If the unit does not pass the unit _ 8 • If the salesman of the patent application scope item 6 described the plural number of salesmen, the sales amount prediction device, of which 舻 媸 μ relative sales: has been obtained during the period Passed units of sales performance; Clever mouth ratio and relative sales ratio after passing the flat period (“Per unit period”: π?): Scoop sales forecast for sales forecast. In the period, the above-mentioned 'Among them, the above 9 orders are like the second and second special :: 7 items on the scope of the lock sale ^ ^ 2 Q 3 is based on the unit period just passed. After the Ϊ: patent scope item 8 sales Pre Vi 11. If applied:! = Uses the unit period that has just passed. The sales forecasting device for item 7 of the patent scope 200300536 ''? Factory :: 4 VI. Application for patent scope The unit period is a unit period. 1-2. The amount of sales in Item II of Patent II, No. 8 is the pre-elapsed unit period, which uses a plurality of units, units, and units. Among them, 13. If the above-mentioned units in Items 7 to 12 of the scope of patent application ^ Among them, regarding the above-mentioned lock sales of the unit-temple period, the estimated lock sales volume and sales of the unit will be used to estimate whether there is insufficient stock. 断 Break: 1 Xiaoji, the above sales performance is corrected based on the purchase. When insufficient, use prediction. / 4 Achievements in sales to achieve sales 14. If any of the items in the scope of the patent application 3 to 12, any of the above, the above-mentioned predetermined unit period is' locked sales forecast for the installation month or year. 'D: Month is% tick, time zone, day, week, 15. If the sales amount corresponding to item 1 of the patent application scope corresponds to the time period corresponding to the time period for which the forecast is made, among which is the previous or previous-day Previous period. The rule of the month-years ago, the first one of the 5 meanings: Γϊ The sales forecast device of the 2nd patent scope, a φ ρ with · 以 I as the corresponding period before the forecast—years are two, two, and the next is Or the previous period. One month, one month, 'where, the above time or climate', where the above time or climate ::, 1 'The similarity of the sales forecast device of the 1st patent scope includes the day, month, week, and day of the week.对应 ΐ ΐ 范围 范围 范围 范围 范围 范围 范围 范围 范围 范围 范围 范围 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 预测 共 预测 共 共 The commonality includes Pei Zhi, including day, month, week, day of week, 19 r- 哕-— 6. Applicants for patent scope, among them, forecast. 2 0. If you apply for a special device, of which the sales performance of the former group during the sales period 21 · If you apply for a special device, of which, the predetermined value is equal to the sales target of the sales transaction during the period 2 2 · If you apply for a special device After the sale, after the purchase is out 2 3 · If the application is calculated after the sale, the next sale is under the purchase 24. If the application is for the specified unit, the deviation period is required 2 5. If the application is for the specified unit, the deviation period is required 2 6 · If you apply for the use of the compound interest range together or later or later, the profit range will be based on the actual sales results in the previous sales range. The range period setting also has a short-term profit range corresponding to the sales period of ^ a. The sales results are used to perform sales lead items 1 to 12, and any of the six-item sales forecast for the forecast will be performed on both sides. ^ The sales forecast of the group to which the barter object belongs. For the rest of the sales period-Items 1 to 12 of which are sales forecasts for h, +, &a; ^ + cooking items.士 展 斜 #% s μ is the change in the overall sales of the entire period, like the above corresponding measurement of the group. Based on the sales forecast of the first item in each unit period, calculate the transaction pair /, where, based on. I must purchase the amount, then the sales forecast of item 2 to calculate the transaction object ^, where, based on. The required purchase amount of the object, however, the sales of item 2 and item 2 are set in terms of ratio, which is proportional to the eighth lieutenant general of the transaction object. The next sales forecast from the beginning to the next delivery is item No. 23, "Yi Zhi, JL φ, and Jiang Bu. 1Sales forecast from the beginning of delivery to delivery No. 24, Pei Zhi, of which, Page 35, 200, 300, 536 6. For the required amount of time required to apply for patents, the amount required is based on the above. 2 7. If you are applying for the purchase amount of a unit of time that is required to apply for a specific amount, then 2 8. If you are applying for a special In the future, the unit sales of the unit will be determined based on the predicted sales measurement inventory 2 9 · If applying for the above-mentioned transaction object 3 0 · If applying for the combination of the above-mentioned transaction object to be grouped 31 · If applying for the above-mentioned grouping 3 2. — A formula for predicting the amount of sales, to obtain the sales volume of each transaction based on the above-mentioned exchange target. The purchase amount in consideration of the above-mentioned deviation period is considered, based on future orders: future unit locks, and sales amount predictions; : The required sales profit forecasting device for the 25th item of the profit margin for purchase, calculated as taking the above deviation period into consideration; above = necessary purchase amount, in the future = based on the above sales forecast-Zhongsuo The required purchase amount for the pre-T period of the sales of item 26 or 27 of Biwanwei is the total amount of the order and the forecast amount of the predetermined number earlier than the unit period. Early The period determines the above-mentioned purchase volume. P is based on the 1st to 12th items of the range of interest # ^ Jaw prediction devices are sold for seasonal reasons. Among them, the range of interest is the first! ^ To 12 items are classified according to their attributes ^ _ \ College / test device, where grouping. The attributes and the above-mentioned selling period are the eclipses of items 1 to 12 in the profit range ~ ^ ^ ^ ^ M t ft ^ ^ 5 ^ t 5 ′ This is the beginning of the period: //: ’. The counties deal with the following: the sales performance of the object of the sale of the banned sales; the sales performance, at least one indicator of each county's post and period :; grouping 'as each transaction Page 36 6. Scope of patent application 6. Scope of patent application for which the forecasted transaction object is ... For the determination of the 'period' to carry out this, in the past correspondence of sex, calculate τ based on the basis: : The period of this transaction group is: 33.-ff λ. The sales rhyme percentage of the sub-images, and the sales performance is the shameless Z-recording media. 34 · A program, which is like applying for a special quota forecasting program, using the light weight of the 32nd item of the electric drying park. For: 夂: The computer will enter the quota forecasting device, and will change each parent's target. The processing of the index transactions is based on the record department of the participation period, so that it can be built in Jishu; at least as one, 口 has recorded the sales performance of each group based on predictions Which group does the elephant belong to; and, when the elephant is sold, to determine the transaction pair, for the forecasted transaction, the period of $ line prediction is similar, Jing: The above-mentioned record department obtains the basis of elephant group The sales of the group are real, and the error ^〗 corresponds to the predicted sales value of the transaction object in the period. Ϋ Based on the performance of the maxillary jaw, calculate the 35 · a recording medium with 36.-locking prediction methods, using ΐ; "; the program around item 34 predicts future sales, electricity, based on the past Its sales performance is characterized by: Obtaining the sales performance of each transaction object; Measured: The estimated sales period of the transaction object to determine the progress. H Whether the easy object corresponds to any one of the existing transaction daggers: 200300536 six The scope of the patent application and the transaction target of the pre-source j, based on the sales performance of the transaction target in the past period similar to the forecast period, are used to calculate the sales target value of the transaction target. Page 38