WO2022265005A1 - 見積推定装置、見積推定方法および見積推定プログラム - Google Patents
見積推定装置、見積推定方法および見積推定プログラム Download PDFInfo
- Publication number
- WO2022265005A1 WO2022265005A1 PCT/JP2022/023782 JP2022023782W WO2022265005A1 WO 2022265005 A1 WO2022265005 A1 WO 2022265005A1 JP 2022023782 W JP2022023782 W JP 2022023782W WO 2022265005 A1 WO2022265005 A1 WO 2022265005A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- unit price
- estimation
- estimated
- calculation
- directly related
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims description 11
- 238000004364 calculation method Methods 0.000 claims abstract description 43
- 238000013528 artificial neural network Methods 0.000 claims abstract description 14
- 238000004458 analytical method Methods 0.000 claims description 32
- 230000004044 response Effects 0.000 claims description 17
- 238000003860 storage Methods 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 description 22
- 239000000463 material Substances 0.000 description 20
- 238000010586 diagram Methods 0.000 description 18
- 238000013068 supply chain management Methods 0.000 description 15
- 238000007726 management method Methods 0.000 description 10
- 238000004040 coloring Methods 0.000 description 7
- 238000000465 moulding Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000015556 catabolic process Effects 0.000 description 5
- 239000003550 marker Substances 0.000 description 5
- 230000009467 reduction Effects 0.000 description 5
- 230000002159 abnormal effect Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- -1 magnification Substances 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 239000011347 resin Substances 0.000 description 1
- 229920005989 resin Polymers 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0206—Price or cost determination based on market factors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0283—Price estimation or determination
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Definitions
- Embodiments of the present invention relate to an estimate estimation device, an estimate estimation method, and an estimate estimation program.
- SCM Supply chain management
- a supply chain is a flow of procurement of parts, etc., as one supply chain.
- a supply chain defines the relationship between a purchaser (buyer) and a seller (supplier).
- the SCM system manages the supply of products such as parts in the supply chain.
- the supplier uses the SCM system to send the quotation reply to the buyer.
- An object of the embodiments is to provide an estimate estimation device, an estimate estimation method, and an estimate estimation program that can easily grasp an abnormal estimate.
- the estimate estimating device of the embodiment uses data directly related to the calculation of the estimated unit price of the estimate reply, data not directly related to the calculation of the estimated unit price of the estimate reply, and the estimated unit price of the estimate reply as learning data, and uses a neural network. and an estimation unit for estimating the unit price of the product/part by using the estimation model generated by the estimation model generation unit.
- FIG. 1 is a configuration diagram showing an example of the configuration of a supply chain management system of one embodiment;
- FIG. It is a schematic diagram for explaining a supply chain.
- 3 is a block diagram showing the configuration of programs and data executed in a server;
- FIG. 10 is a diagram showing an example of a quotation reply sent from a supplier to a buyer;
- FIG. 10 is a diagram showing an example of input items to be input to the neural network from the items of quotation details of the quotation reply;
- FIG. 10 is a diagram for explaining an example of the flow of estimation model generation processing and similar product search processing; It is a figure which shows the structural example of a neural network.
- FIG. 7 is a flowchart illustrating an example of the flow of estimation model generation processing; It is a flowchart which shows an example of the flow of estimation processing of estimated unit price. It is a figure which shows an example of the analysis report which shows the relationship between an estimated unit price and an estimated unit price. It is a figure which shows an example of the importance of the item of quotation breakdown.
- FIG. 10 is a diagram showing an example of a similar product search screen when searching for similar products using importance. 10 is a flowchart showing an example of the flow of similar product search processing using importance. It is a figure which shows an example of a standard unit price table.
- FIG. 1 is a configuration diagram showing an example of the configuration of a supply chain management system according to one embodiment.
- the supply chain management system 1 includes a server 2 as a quotation management device, multiple terminals 3 for multiple buyers, multiple terminals 4 for multiple suppliers, and a communication network 5 .
- the server 2 and a plurality of terminals 3 and 4 are communicably connected to each other via a network 5.
- Network 5 is the Internet here.
- Each terminal 3, 4 has an input device and a display device 3a.
- the input device is a keyboard, mouse, etc.
- the display device 3a is a monitor.
- only one terminal 3 has a display device 3a.
- the server 2 has a processor 11 and a storage device 12.
- the storage device 12 stores various software programs for a supply chain management system, which will be described later, and various information.
- the supply chain management system 1 is a system for managing the supply chain of products such as parts purchased or supplied by multiple buyers and multiple suppliers.
- the server 2 can transmit supply chain risk information to the terminals 3 and 4 via the network 5 in response to requests from the terminals 3 and 4 via the network 5 .
- Each buyer can use the supply chain management system 1 to manage their own supply chain. Therefore, each buyer can access the server 2 from his terminal 3 and register his own supply chain information.
- Each buyer may register all of its own supply chain information, each supplier may register its own supply chain information, and a primary supplier may register secondary and subsequent supplier information. .
- each buyer can access the server 2 from his/her own terminal 3 and register information (business partner basic information (including business partner risk information RI)) about each supplier.
- business partner basic information including business partner risk information RI
- each buyer and each supplier accesses the server 2 via the network 5 using their own terminals 3 and 4, they can input, display, and output data on a screen using a browser. Access to the server 2 from each of the terminals 3 and 4 becomes possible after various authentications.
- FIG. 2 is a schematic diagram for explaining the supply chain.
- FIG. 2 shows an example supply chain for parts X and Y.
- FIG. FIG. 2 shows a case where a certain buyer purchases part X from supplier A and part Y from supplier D, and manufactures and sells his own products.
- suppliers A and D are primary suppliers of parts X and Y, respectively.
- suppliers A and D purchase parts x1 and y1 from suppliers B and E in order to manufacture and sell parts X and Y, respectively.
- suppliers B and E purchase parts x2 and y2 from suppliers C and F in order to manufacture and sell parts x1 and y1, respectively.
- suppliers C and F purchase parts x3 and y3 from other suppliers in order to manufacture and sell parts x2 and y2, respectively. That is, the supply chain includes multiple tiers. Therefore, each supplier can also be a buyer.
- the supply chain may branch into multiple suppliers.
- the buyer may receive parts supplied from two suppliers regarding part X
- supplier D may also receive parts supplied from two suppliers.
- FIG. 3 is a block diagram showing the configuration of programs and data executed on the server.
- the server 2 has a software program that manages transactions between buyers and suppliers by SRM (Supplier Relationship Management).
- Supply chain management system 1 stores a software program for SRM in storage device 12 .
- FIG. 3 for SRM in supply chain management, procurement analysis part PA, BCP (business continuity planning) management part, general document exchange part, electronic quotation part, workflow management part, document management part, customer basic information part,
- the estimation model generating unit MG and the estimating unit EP are shown, there are other various processing units.
- it also has a portal program for access management to the supply chain management system 1 .
- the buyer portal is a processing unit for the buyer to access the server 2, and performs various authentication processes. Once properly authenticated, the buyer can use server 2 .
- the supplier portal is a processing unit for the supplier to access the server 2, and performs various authentication processes. Once properly authenticated, the supplier can use server 2 .
- the procurement analysis part PA, BCP management part, etc. are stored in the storage device 12 as software programs, and can be read out and executed by the processor 11 when necessary.
- the Procurement Analysis Department PA analyzes each buyer's procurement to suppliers.
- Procurement analysis department PA can generate analysis reports such as the number of estimates, estimate results, various evaluations, etc., for each supplier related to itself, based on estimate history information ERI, for example.
- the BCP management department collects information on the latitude/longitude of each supplier's bases that make up the supply chain, and identifies the bases that exist within the affected area in the event of an emergency such as a disaster.
- the general-purpose document exchange department exchanges documents between buyers and suppliers.
- the electronic quotation department manages requests for quotations to suppliers and quotation responses from suppliers. Buyers can use electronic quoting to request product/component quotes from one or more suppliers. The request for quotation is sent to the one or more suppliers, and the supplier can send a quotation reply to the buyer.
- the workflow management department manages the workflow of various processes between buyers and suppliers.
- the document management department manages documents created by buyers and documents received from suppliers.
- the Business Partner Basic Information Department registers and manages basic information about primary suppliers (capital, president, business partner risk information RI, etc.) in the storage device 12 as business partner information BAI. If the buyer cannot register information about all suppliers, primary suppliers may be allowed to register basic information about secondary and lower suppliers.
- the business partner risk information RI included in the business partner information includes the quantified degree of risk r regarding the supplier's disaster countermeasures.
- the degree of risk r is a value quantified based on a predetermined evaluation criterion.
- the degree of risk r has four levels.
- Level 1 has the lowest degree of risk r, for example, when sufficient disaster countermeasures are taken and sufficient product inventory is secured.
- Level 4 has the highest degree of risk r, and is a case where, for example, disaster countermeasures are inadequate.
- the degree of risk r for secondary and lower suppliers is determined by the primary supplier.
- the estimation model generation unit MG reads estimation responses such as estimation history information ERI, learns by deep learning, and generates an estimation model.
- the estimation model generation unit MG performs learning by deep learning for each type of product/part and processing, such as resin molding processing, cutting processing, and assembly processing, to generate an estimation model.
- the estimation model generating unit MG stores the multiple generated estimation models as an estimation model group M in the storage device 12 .
- the estimation part EP When the supplier sends a new quotation reply based on the request for quotation from the buyer, the estimation part EP inputs the quotation reply into the estimation model of the corresponding product/part in the estimation model group M, and estimates the unit price of the product/part. do.
- the estimation unit EP can calculate the difference between the estimated unit price described in the estimate reply and the estimated unit price estimated by AI, and display the calculation result on the display device 3a.
- the storage device 12 of the server 2 also stores various types of information. In FIG. 3, only supplier information BAI, supply chain information SCI, estimate history information ERI, and estimation model group M are shown.
- the supplier information BAI is, as described above, basic information about primary suppliers (capital, president, supplier risk information RI, etc.).
- the supply chain information SCI is base information such as primary suppliers, secondary suppliers, and tertiary suppliers for each delivered item (parts, products, etc.) in the supply chain.
- Base information such as primary suppliers, secondary suppliers, and tertiary suppliers is registered.
- Each site information includes a site name and location information.
- Each base is a place where a factory such as a subcontractor is located.
- the location information includes latitude/longitude information. Location information for secondary suppliers, tertiary suppliers, etc. is registered by the primary supplier.
- the quotation history information ERI is information such as quotation requests, quotation responses, and quotation results.
- the estimated model group M includes information on a plurality of estimated models generated by the estimated model generator MG.
- FIG. 4 is a diagram showing an example of a quotation reply sent from a supplier to a buyer.
- the supplier fills in the quotation reply E in FIG. 4 and then sends it to the buyer via the electronic quotation section.
- the buyer carefully examines the estimate reply E, and confirms whether there is any abnormality in the estimated unit price (estimated amount) of the product/parts.
- the estimate reply E is an estimate related to the molded product, and includes entry fields for material costs, coloring costs, molding processing costs, etc.
- the estimate reply E includes not only the fields for filling in the material cost, coloring cost, and molding processing cost, but also the material information including items such as the material manufacturer, and the unit price details including items such as administrative expenses and profits, etc. (not shown). has columns.
- Material cost items include material code, material name, specifications, quantity (g), material unit price, and loss (%).
- the item of coloring cost includes coloring material, magnification, material name, quantity (g), coloring unit price, and loss (%).
- Items of molding processing cost include molding machine size, quantity, model unit price/day, number of shots (times/day), and loss (%).
- the items for which numerical data are entered and are directly related to the calculation of the estimated unit price are item A1 quantity (g), item A2 material unit price, and item A3 loss (%).
- the items for which numerical data are not input and which are not directly related to the calculation of the estimated unit price are the material code of item B1, the material name of item B2, and the specification of item B3.
- the items directly related to the calculation of the estimated unit price are the quantity (g), the coloring unit price, and the loss (%).
- the material name is the material name.
- the items directly related to the calculation of the estimated unit price are the size of the molding machine, the quantity, the model unit price/day, the number of shots (times/day), and the loss (%). There are no items that are not directly related to the calculation.
- FIG. 5 is a diagram showing an example of input items to be input to the neural network from items of quotation details of the quotation reply.
- item A is the item name to which numerical data such as items A1, A2, and A3 of estimate reply E are input
- item B is non-numerical data such as items B1, B2, and B3 of estimate reply E (pull-down data) is the name of the item to be entered. That is, item A is an item name in which numerical data such as "quantity (g)", “material unit price” and “loss (%)" directly related to the calculation of the estimated unit price are entered.
- This is an item name in which non-numerical data such as "material code", "material name” and “specification” which are not directly related to unit price calculation are input.
- the item name of item A and numerical data corresponding to the item name, and the item name of item B and non-numerical data corresponding to the item name are provided to the input layer of the neural network NN described later.
- the data not directly related to the calculation of the estimated unit price (the item name of item A and the input data (numerical data) corresponding to the item name)
- An estimation model is generated using the item name of item B and the input data (non-numerical data) corresponding to the item name.
- the item names of item A that are directly related to the calculation of the estimated unit price and the input data (first data set) corresponding to the item names, and the items that are not directly related to the calculation of the estimated unit price are given to the input layer, and the estimated unit price is given to the output layer to generate an estimation model. Since the item name of item A, the item name of item B, and the input data corresponding to the item name of item B are non-numeric data, preprocessing for numerical conversion is performed.
- FIG. 6 is a diagram for explaining an example of the flow of estimation model generation processing and similar product search processing.
- the procurement analysis unit PA reads the quotation history information ERI and preprocesses the learning data. Specifically, the procurement analysis unit PA extracts the item names of item A that are directly related to the calculation of the estimated unit price and numerical data corresponding to the item names, and extracts the item names of item B that are not directly related to the calculation of the estimated unit price. and extract non-numeric data corresponding to item names. Procurement analysis unit PA then performs preprocessing for digitizing the input data corresponding to the item names of item A, item B, and item B, which are non-numeric data.
- the estimation model generation unit MG generates data corresponding to the item name of item A, which is numerical data, the item name of item A converted to numerical data by preprocessing, the item name of item B, and the item name of item B. and data corresponding to , are given to the input layer, and the estimated unit price is given to the output layer, learning processing is performed, and an estimation model is generated.
- the procurement analysis department PA generates the analysis report shown in FIG. 10 based on the quotation history information ERI. Further, the procurement analysis unit PA acquires importance items and importance coefficients shown in FIG. 11 to be described later when the estimation model generation unit MG generates an estimation model.
- the procurement analysis part PA can search for similar products based on the similarity of items with high importance coefficients and the similarity of importance coefficients.
- FIG. 7 is a diagram showing a configuration example of a neural network.
- the estimation models that make up the estimation model group M are generated using the neural network NN shown in FIG.
- the neural network NN has an input layer 31, a hidden layer 32, and an output layer 33.
- the input layer 31 stores the item name of item A directly related to the estimated unit price and the data corresponding to the item name, and the item name and item name of item B not directly related to the estimated unit price. It has as many input units 31a indicated by circles as the number of corresponding data elements.
- the hidden layer 32 has a multi-layer structure including multiple hidden layers 32a.
- the output layer 33 has one output unit 33a, and the one output unit 33a is given the estimated unit price of the estimated breakdown items to generate an estimated model.
- An estimation model is generated for each product/part, and a plurality of generated estimation models are stored in the storage device 12 as an estimation model group.
- FIG. 8 is a flowchart illustrating an example of the flow of estimation model generation processing.
- Procurement analysis unit PA reads quotation history information ERI (S1).
- Procurement analysis part PA preprocesses learning data from quotation reply E in quotation history information ERI (S2).
- the input data corresponding to the item name of item A, the item name of item B, and the item name of item B, which are non-numeric data are converted into numerical values.
- the estimation model generation unit MG uses deep learning to learn items of the quotation details and generate an estimation model (S3).
- the item of the quotation details corresponds to the item name of item A that is directly related to the calculation of the estimated unit price and the input data (numerical data) corresponding to the item name, and the item name of item B that is not directly related to the calculation of the estimated unit price and the item name.
- the estimation model generator MG saves the estimation model in the storage device 12 (S4), and terminates the process.
- FIG. 9 is a flow chart showing an example of the flow of an estimated unit price estimation process.
- the estimation unit EP inputs the estimate reply E to the estimation model (S11). Specifically, the estimating part EP stores the item name of the item A directly related to the calculation of the estimated unit price among the items of the estimate breakdown of the estimate reply E, the input data (numerical data) corresponding to the item name, and the estimated unit price. Input data (non-numerical data) corresponding to the item name of item B and the item name, which are not directly related to the calculation of , into the estimation model. As described above, since the input data corresponding to the item name of item A, the item name of item B, and the item name of item B are non-numeric data, preprocessing for numerical conversion is performed.
- the preprocessing is performed by the procurement analysis part PA.
- the estimating unit EP estimates the estimated unit price of the product or part (S12) and terminates the process.
- Input data (numerical data) corresponding to the item name of item A that is directly related to the calculation of the estimated unit price and input data (numerical data) corresponding to the item name of item B that is not directly related to the calculation of the estimated unit price and input data (non- Numerical data) is input to the estimation model, and an estimated unit price (estimated unit price) estimated by the estimation model is output from the estimation unit EP.
- the estimated unit price estimated in this way is used for generating an analysis report by the procurement analysis part PA.
- the estimation model generation unit MG generates the item names of the item A directly related to the calculation of the estimated unit price, the input data (numerical data) corresponding to the item names, and the item B not directly related to the calculation of the estimated unit price.
- An estimation model is generated using item names and input data (non-numeric data) corresponding to the item names.
- the estimation unit EP estimates the estimated unit price using the estimated model thus generated.
- the estimating unit EP can estimate the estimated unit price taking into consideration items that would otherwise be overlooked by a veteran buyer (the item name of item B and the input data corresponding to the item name, which are not directly related to the calculation of the estimated unit price). This allows buyers to easily spot unusual quotes.
- FIG. 10 is a diagram showing an example of an analysis report showing the relationship between estimated unit prices and estimated unit prices.
- the horizontal axis indicates the estimated unit price estimated by the estimation model
- the vertical axis indicates the estimated unit price estimated by the estimation reply E.
- the procurement analysis part PA can generate an analysis report based on the estimated unit price and the estimated unit price estimated by the estimation part EP using the estimation model.
- the relationship between the estimated unit price of the estimate reply E and the estimated unit price obtained by the estimation model is displayed by the marker MK.
- the straight line L is a line that connects points where the estimated unit price and the estimated unit price are the same.
- the marker MK above the straight line L indicates that the estimated unit price is higher than the estimated unit price
- the marker MK below the straight line L indicates that the estimated unit price is lower than the estimated unit price. That is, the greater the distance from the straight line L to the marker MK, the greater the difference between the estimated unit price and the estimated unit price.
- the priority cost reduction areas in the analysis report are products/parts with high unit prices and areas where the estimated unit price is higher than the estimated unit price. Therefore, by preferentially lowering the unit prices of the products/parts in the priority cost reduction area, a large cost reduction can be achieved.
- the product/part is automatically adopted without manually checking the estimate reply E. area.
- FIG. 11 is a diagram showing an example of the importance of items in the quotation details.
- the estimation model generation unit MG When the estimation model generation unit MG generates the estimation model, it calculates the importance (importance coefficient) indicating which item (variable) influences the estimation of the estimated price and to what extent.
- the procurement analysis unit PA displays, for example, 15 items with high importance (importance coefficient) calculated by the estimation model generation unit MG on the display device 3a. As a result, the top 15 important items (variables) that contribute to the calculation of the estimated unit price are displayed on the display device 3a.
- the buyer can grasp the highly important items (variables) that contribute to the unit price of the product/parts, which helps the buyer to check which items should be focused on when assessing the quotation reply E. .
- highly important items (variables) that contribute to the unit price of products/parts buyers can obtain useful information for price negotiations and continuously reduce costs.
- the procurement analysis unit PA can search for similar products from the quotation history information ERI of past quotations using the importance (importance coefficient) calculated by the estimation model generation unit MG.
- FIG. 12 is a diagram showing an example of a similar product search screen when searching for similar products using importance.
- the similar product search screen 40 has input fields 41, 42 and 43 in which importance coefficients A, B and C are input (selected).
- the order of importance coefficients A, B, and C is the order of importance.
- the similar product search screen 40 also has an input field 44 in which the upper limit value of the importance coefficient A is input, and an input field 45 in which the lower limit value of the importance coefficient A is input.
- An input field 41 for inputting (selecting) the importance coefficient A is provided with a selection button 46 for selecting the importance coefficient.
- Importance coefficients B and C have the same configuration.
- a list of importance coefficients is displayed.
- the user can select a desired importance coefficient from the displayed list of importance coefficients.
- the user can also input the upper limit value of the selected importance coefficient A in the input field 44 and the lower limit value in the input field 45 .
- the [material cost] quantity (g) is selected as the importance coefficient A, and 24 is input as the upper limit and 12 as the lower limit.
- the similar product search screen 40 also has a search button 47 .
- the user selects the importance coefficients A, B, and C, inputs the upper limit value and lower limit value of each coefficient, and presses the search button 47 to search for similar products from the quotation history information ERI that has been quoted in the past. can be searched.
- the selected importance coefficients are not limited to the three importance coefficients A, B, and C, and may be one, two, or four or more. Also, it is not necessary to input both the upper limit value and the lower limit value of the selected importance coefficient, and either one may be entered. Also, if the selected importance coefficient is item B of non-numeric data, there is no need to enter the upper and lower limits.
- the user used the similar product search screen 40 to arbitrarily set the order of importance (coefficients) and the range of the upper and lower limits of each coefficient, thereby adjusting the search conditions and making estimates in the past.
- a similar product can be searched from the quotation history information ERI.
- FIG. 13 is a flowchart showing an example of the flow of similar product search processing using importance.
- the procurement analysis unit PA acquires the items with high importance and the importance coefficients calculated by the estimation model generation unit MG (S21).
- Procurement analysis department PA selects quotation replies E with similar importance coefficients and the order of importance set on the similar product search screen 40 from past quotation replies E stored as quotation history information ERI. Extract (S22).
- Procurement analysis part PA outputs the extracted product/part of quotation reply E to display device 3a as a similar product (S23), and ends the process.
- FIG. 14 is a diagram showing an example of the standard unit price table.
- the estimation model generation unit MG selects items that serve as the basis for estimating the estimated unit price when learning the item of the quotation breakdown by the neural network NN and generating the estimation model, and sets the standard unit price for each of the selected items. calculate. As shown in FIG. 13, the calculated standard unit price is associated with the quotation breakdown name, item name, unit/reference condition, and currency, and stored in the storage device 12 as a standard unit price table.
- the buyer can obtain the standard unit price of item B, which is not directly related to the estimated unit price, so that the buyer can ascertain whether the estimated unit price of the estimate reply E is normal.
- each step in the flow charts in this specification may be executed in a different order, may be executed at the same time, or may be executed in a different order for each execution, as long as it does not contradict its nature.
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Data Mining & Analysis (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
Description
図1は、一実施形態のサプライチェーン管理システムの構成の一例を示す構成図である。サプライチェーン管理システム1は、見積管理装置としてのサーバ2と、複数のバイヤー用の複数の端末3と、複数のサプライヤー用の複数の端末4と、通信用のネットワーク5を含む。
バイヤーが電子見積部を経由してサプライヤーに見積依頼を行うと、サプライヤーが図4の見積回答Eを記入後、電子見積部を経由してバイヤーに送付する。バイヤーは、見積回答Eの精査を行い、製品/部品の見積単価(見積金額)に異常がないかを確認する。
図5において、項目Aは見積回答Eの項目A1、A2、A3等の数値データが入力される項目名であり、項目Bは見積回答Eの項目B1、B2、B3等の非数値データ(プルダウンデータ)が入力される項目名である。すなわち、項目Aは、見積単価の算出に直接関係する「数量(g)」、「素材単価」及び「ロス(%)」等の数値データが入力される項目名であり、項目Bは、見積単価の算出に直接関係しない「素材コード」、「素材名称」及び「仕様」等の非数値データが入力される項目名である。項目Aの項目名および項目名に対応する数値データと、項目Bの項目名および項目名に対応する非数値データが後述するニューラルネットワークNNの入力層に与えられる。
調達分析部PAは、見積履歴情報ERIを読み込んで学習データの前処理を行う。具体的には、調達分析部PAは、見積単価の算出に直接関係する項目Aの項目名および項目名に対応する数値データを抽出するとともに、見積単価の算出に直接関係しない項目Bの項目名および項目名に対応する非数値データを抽出する。そして、調達分析部PAは、非数値データである項目Aの項目名、項目Bの項目名、および、項目Bの項目名に対応する入力データを数値化する前処理を行う。推定モデル生成部MGは、数値データである項目Aの項目名に対応するデータと、前処理によって数値データに変換された項目Aの項目名、項目Bの項目名、および、項目Bの項目名に対応するデータとを入力層に与え、見積単価を出力層に与えて学習処理を行い、推定モデルを生成する。
出力層33は、1つの出力ユニット33aを有し、1つの出力ユニット33aに見積内訳項目の見積単価が与えられ、推定モデルが生成される。推定モデルは、製品/部品毎に生成され、生成された複数の推定モデルが推定モデル群として記憶装置12に記憶される。
調達分析部PAが見積履歴情報ERIを読み込む(S1)。調達分析部PAが見積履歴情報ERIの見積回答Eから学習データの前処理を行う(S2)。前処理では、上述したように、非数値データである項目Aの項目名、項目Bの項目名、および、項目Bの項目名に対応する入力データを数値化する処理が行われる。
推定部EPが見積回答Eを推定モデルに入力する(S11)。具体的には、推定部EPは、見積回答Eの見積内訳の項目のうち、見積単価の算出に直接関係する項目Aの項目名および項目名に対応する入力データ(数値データ)と、見積単価の算出に直接関係しない項目Bの項目名および項目名に対応する入力データ(非数値データ)とを推定モデルに入力する。上述したように、項目Aの項目名、項目Bの項目名、および、項目Bの項目名に対応する入力データは、非数値データであるため、数値化するための前処理を行う。なお、前処理は、調達分析部PAによって行われる。推定部EPが製品または部品の見積単価を推定し(S12)、処理を終了する。見積単価の算出に直接関係する項目Aの項目名および項目名に対応する入力データ(数値データ)と、見積単価の算出に直接関係しない項目Bの項目名および項目名に対応する入力データ(非数値データ)とを推定モデルに入力することで、推定モデルによって推定された見積単価(推定単価)が推定部EPから出力される。このように推定された推定単価は調達分析部PAによる分析レポートの生成に用いられる。
図10において、横軸は推定モデルによって推定された推定単価を示し、縦軸は見積回答Eで見積された見積単価を示している。調達分析部PAは、見積単価と推定部EPが推定モデルを用いて推定した推定単価とに基づいて分析レポートを生成することができる。具体的には、図10に示すように、見積回答Eの見積単価と、推定モデルによって得られた推定単価との関係をマーカーMKによって表示する。
推定モデル生成部MGが推定モデルを生成する際に、どの項目(変数)が見積価格の推定にどの程度の影響を与えているかを示す重要度(重要度係数)を算出する。図11に示すように、調達分析部PAは、推定モデル生成部MGにより算出された重要度(重要度係数)が大きい、例えば15個の項目を表示装置3aに表示する。この結果、推定単価を算出する際に寄与する上位15個の重要な項目(変数)が表示装置3aに表示される。
類似品検索画面40は、重要度係数A、B及びCが入力(選択)される入力欄41、42及び43を有する。重要度係数A、B及びCの順番は、重要度の高い順番となっている。また、類似品検索画面40は、重要度係数Aの上限値が入力される入力欄44、重要度係数Aの下限値が入力される入力欄45を有する。重要度係数Aが入力(選択)される入力欄41には、重要度係数を選択するための選択ボタン46が設けられている。重要度係数BおよびCについても同様の構成となっている。
調達分析部PAは、推定モデル生成部MGにより算出された重要度の高い項目及び重要度係数を取得する(S21)。調達分析部PAは、見積履歴情報ERIとして記憶されている過去の見積回答Eから、類似品検索画面40で設定された重要度の高い項目の順番、及び、重要度係数が近い見積回答Eを抽出する(S22)。調達分析部PAは、抽出した見積回答Eの製品/部品を類似品として表示装置3aに出力し(S23)、処理を終了する。
Claims (10)
- 見積回答の見積単価の算出に直接関係するデータと、前記見積回答の見積単価の算出に直接関係しないデータと、前記見積回答の見積単価とを学習データとし、ニューラルネットワークを用いて推定モデルを生成する推定モデル生成部と、
前記推定モデル生成部が生成した前記推定モデルを用いて、製品/部品の単価を推定する推定部と、
を有する、見積推定装置。 - 前記見積回答の見積単価の算出に直接関係するデータは、前記見積単価の算出に直接関係する項目の項目名、および、前記見積単価の算出に直接関係する項目の項目名に対応する入力データを含み、
前記見積回答の見積単価の算出に直接関係しないデータは、前記見積単価の算出に直接関係しない項目の項目名、および、前記見積単価の算出に直接関係しない項目の項目名に対応する入力データを含む、請求項1に記載の見積推定装置。 - 前記見積単価の算出に直接関係する項目の項目名、前記見積単価の算出に直接関係しない項目の項目名、および、前記見積単価の算出に直接関係しない項目の項目名に対応する入力データを数値化する前処理を行う分析部を有する、請求項2に記載の見積推定装置。
- 前記分析部は、前記推定部により推定された単価と、前記見積回答の見積単価との関係を示す分析レポートを生成して表示装置に表示する請求項3に記載の見積推定装置。
- 前記分析部は、前記推定モデル生成部が前記推定モデルを生成する際に前記単価を推定するのにどの程度の影響を与えているかを示す重要度係数を算出する請求項3に記載の見積推定装置。
- 前記分析部は、前記重要度係数の高い項目の類似度及び/又は前記重要度係数の類似度によって前記製品/部品に類似した類似品を検索する請求項5に記載の見積推定装置。
- 前記推定モデル生成部は、前記推定モデルを生成するときに前記単価を推定する際の根拠となる項目を選出し、選出した項目についてそれぞれ標準単価を算出する請求項1に記載の見積推定装置。
- 前記推定モデル生成部は、前記製品/部品毎に前記推定モデルを生成し、生成した複数の推定モデルを推定モデル群として記憶装置に記憶する請求項1に記載の見積推定装置。
- 見積回答の見積単価の算出に直接関係するデータと、前記見積回答の見積単価の算出に直接関係しないデータと、前記見積回答の見積単価とを学習データとし、ニューラルネットワークを用いて推定モデルを生成し、
生成した前記推定モデルを用いて、製品/部品の単価を推定する、
見積推定方法。 - コンピュータに、
見積回答の見積単価の算出に直接関係するデータと、前記見積回答の見積単価の算出に直接関係しないデータと、前記見積回答の見積単価とを学習データとし、ニューラルネットワークを用いて推定モデルを生成するステップと、
生成した前記推定モデルを用いて、製品/部品の単価を推定するステップと、
を実行させるための見積推定プログラム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
EP22824993.4A EP4358004A1 (en) | 2021-06-15 | 2022-06-14 | Quotation estimation device, quotation estimation method, and quotation estimation program |
CN202280015045.9A CN116940953A (zh) | 2021-06-15 | 2022-06-14 | 报价推定装置、报价推定方法及报价推定程序 |
US18/469,054 US20240029094A1 (en) | 2021-06-15 | 2023-09-18 | Estimate presumption apparatus, estimate presumption method, and estimate presumption program |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021099649A JP2022191039A (ja) | 2021-06-15 | 2021-06-15 | 見積推定装置、見積推定方法および見積推定プログラム |
JP2021-099649 | 2021-06-15 |
Related Child Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US18/469,054 Continuation US20240029094A1 (en) | 2021-06-15 | 2023-09-18 | Estimate presumption apparatus, estimate presumption method, and estimate presumption program |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022265005A1 true WO2022265005A1 (ja) | 2022-12-22 |
Family
ID=84526523
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2022/023782 WO2022265005A1 (ja) | 2021-06-15 | 2022-06-14 | 見積推定装置、見積推定方法および見積推定プログラム |
Country Status (5)
Country | Link |
---|---|
US (1) | US20240029094A1 (ja) |
EP (1) | EP4358004A1 (ja) |
JP (1) | JP2022191039A (ja) |
CN (1) | CN116940953A (ja) |
WO (1) | WO2022265005A1 (ja) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006195738A (ja) * | 2005-01-13 | 2006-07-27 | Mitsubishi Electric Corp | 電子商取引システム |
JP2007179446A (ja) * | 2005-12-28 | 2007-07-12 | Canon System Solutions Inc | 情報処理装置及びその制御方法、並びにプログラム |
JP2019032623A (ja) * | 2017-08-04 | 2019-02-28 | 富士通株式会社 | 部品見積もりプログラム、部品見積もりシステム及び部品見積もり方法 |
JP2021099649A (ja) | 2019-12-20 | 2021-07-01 | ヤフー株式会社 | 決定装置、決定方法及び決定プログラム |
-
2021
- 2021-06-15 JP JP2021099649A patent/JP2022191039A/ja active Pending
-
2022
- 2022-06-14 EP EP22824993.4A patent/EP4358004A1/en active Pending
- 2022-06-14 CN CN202280015045.9A patent/CN116940953A/zh active Pending
- 2022-06-14 WO PCT/JP2022/023782 patent/WO2022265005A1/ja active Application Filing
-
2023
- 2023-09-18 US US18/469,054 patent/US20240029094A1/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006195738A (ja) * | 2005-01-13 | 2006-07-27 | Mitsubishi Electric Corp | 電子商取引システム |
JP2007179446A (ja) * | 2005-12-28 | 2007-07-12 | Canon System Solutions Inc | 情報処理装置及びその制御方法、並びにプログラム |
JP2019032623A (ja) * | 2017-08-04 | 2019-02-28 | 富士通株式会社 | 部品見積もりプログラム、部品見積もりシステム及び部品見積もり方法 |
JP2021099649A (ja) | 2019-12-20 | 2021-07-01 | ヤフー株式会社 | 決定装置、決定方法及び決定プログラム |
Also Published As
Publication number | Publication date |
---|---|
JP2022191039A (ja) | 2022-12-27 |
CN116940953A (zh) | 2023-10-24 |
US20240029094A1 (en) | 2024-01-25 |
EP4358004A1 (en) | 2024-04-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Biscaia et al. | Assessing customer satisfaction and loyalty in the retail sector | |
Lee et al. | A portfolio model for component purchasing strategy and the case study of two South Korean elevator manufacturers | |
Bottani et al. | A fuzzy multi-attribute framework for supplier selection in an e-procurement environment | |
Şen et al. | Pre-selection of suppliers through an integrated fuzzy analytic hierarchy process and max-min methodology | |
US9037733B2 (en) | System and method for enabling product development | |
US20070219877A1 (en) | Method and system for managing home shopper data | |
JPWO2006004132A1 (ja) | 企業評価寄与因子及び/又は指標特定装置、特定プログラム並びに特定方法 | |
WO2008077074A2 (en) | Determining and presenting product market prices | |
JP2013045264A (ja) | クラウドマーケットプレイスシステム | |
KR101448290B1 (ko) | 전문가의 낙찰 예측 분포도를 이용하는 입찰 컨설팅 시스템 및 그 입찰 컨설팅 방법 | |
KR20190088134A (ko) | 인공지능을 이용한 시설관리 서비스 | |
JP3544510B2 (ja) | 取引先総合力判定システム | |
US8204772B2 (en) | Customer service experience comparative landscape tool | |
WO2022265005A1 (ja) | 見積推定装置、見積推定方法および見積推定プログラム | |
JP5358038B2 (ja) | 商品管理サーバ及び商品管理方法 | |
KR100892130B1 (ko) | 온라인 주문서 평가 시스템 및 그 방법 | |
US11430033B2 (en) | Methods and systems of utilizing machine learning to provide trust scores in an online automobile marketplace | |
CN113744024A (zh) | 一种商家匹配方法、装置、计算机设备及存储介质 | |
Das et al. | Models for supply chain vendor selection in e-markets | |
JP2002117131A (ja) | 格付け方法及びこれを用いた電子商取引方法 | |
KR20190084145A (ko) | 온라인 특허 경매 시스템을 이용한 유망 기술 발굴 인공지능 시스템 | |
KR102572171B1 (ko) | 산업재 및 소비재 통합 구매 서비스 제공 시스템 | |
KR102541691B1 (ko) | 지역 기반 상품정보 공유를 위한 참여자 보상정보관리컴퓨터 및 그 방법 | |
Vrat et al. | Source selection, performance rating, and development | |
Toste et al. | A critical overview of the RFQ process of a Global Company |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22824993 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202280015045.9 Country of ref document: CN |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2022824993 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
ENP | Entry into the national phase |
Ref document number: 2022824993 Country of ref document: EP Effective date: 20240115 |