Derivation system, derivation device, and derivation program

Abstract

The present invention provides a derivation system, derivation device, and derivation program that can improve the accuracy of predictions regarding the use of transport means when deriving transport plans for transporting materials necessary for building construction. [Solution] The derivation system includes a trained model that stores the relationship between actual property information, which is property information of completed buildings, and actual transportation records by transportation means. The trained model is generated by machine learning using selected elements, which are pre-selected from multiple elements included in the actual property information based on their relationship with actual transportation records, as input data, and the corresponding transportation records as labeling data. Based on the acquired target property information, the derivation system refers to the trained model and derives a transportation plan for transportation means that transports the materials necessary for the construction of the building indicated in the target property information (S201~S217).

Description

Technical Field

[0001] This application discloses a derivation system for deriving a conveyance plan of a conveyance means for conveying materials necessary for building a building, a derivation device used in such a derivation system, and a derivation program for realizing such a derivation device.

Background Art

[0002] Every day, many trucks are operated across the country to transport housing materials from factories to construction sites. When formulating a conveyance plan for the operation of trucks, for example, after receiving a conveyance instruction at the factory, the person in charge uses a prediction formula using coefficients empirically grasped by each factory for the property information and the member information after member deployment, and practices such as predicting the number of trucks one month ahead are carried out.

[0003] For example, in Patent Document 1, a method for creating a delivery plan for building materials is proposed, which analyzes processes from building design data and formulates a delivery schedule for delivery to a material storage yard.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] However, at factories, in consideration of various problems such as the difficulty of securing trucks and the 2024 problem, it is required to accurately predict future demand. Against such requirements, predictions based on prediction formulas have problems such as the predictable period being about one month and difficulty in long-term prediction. Also, predictions using AI (Artificial Intelligence) technology as described in Patent Document 1 have the problem of improving prediction accuracy.

[0006] The derivation system disclosed in this application was developed in view of these circumstances, and is expected to be able to handle long-term predictions and aims to provide a derivation system that can improve prediction accuracy.

[0007] Furthermore, this application also aims to provide a derivation device used in the aforementioned derivation system and a derivation program that realizes such a derivation device. [Means for solving the problem]

[0008] To solve the above problems, the derivation system disclosed in this application is a derivation system using a derivation device for deriving a transport plan for transporting materials necessary for the construction of a building, comprising: an acquisition means for acquiring target property information, which is property information relating to a building that is the subject of the derivation of the transport plan; and a trained model that stores the relationship between actual property information, which is property information of a completed building, and transport results by the transport means, wherein the trained model is generated by machine learning using training data in which selected elements, which are selected in advance from a plurality of elements included in the actual property information based on their relationship with transport results, are input data, and the corresponding transport results are used as labels, and the derivation device comprises a plan derivation means that, based on the target property information acquired by the acquisition means, refers to the trained model and derives a transport plan for transporting materials necessary for the construction of a building indicated in the target property information.

[0009] Furthermore, the derivation system is characterized in that the trained model selects elements as selected elements whose contribution rate is higher than a predetermined value in a multiple regression analysis using multiple elements included in the actual property information as explanatory variables and transportation performance as the dependent variable.

[0010] Furthermore, in the derivation system, the trained model is generated using the number of transports, transport intervals, and transport capacity of transport records as training data, and the plan derivation means derives a transport plan based on the derivation results of the number of transports, transport intervals, and transport capacity.

[0011] Furthermore, in the derivation system, the trained model includes a transport count model generated using the transport count of transport results as training data, a transport interval model generated using the transport interval of transport results as training data, and a transport capacity model generated using the transport capacity of transport results as training data. The plan derivation means derives the transport count based on the target object information acquired by the acquisition means by referring to the transport count model, derives the transport interval based on the target object information acquired by the acquisition means and the derived transport count by referring to the transport interval model, derives the transport capacity based on the target object information acquired by the acquisition means and the derived transport count by referring to the transport capacity model, and derives a transport plan based on the derived transport interval and transport capacity.

[0012] Furthermore, the derivation system is characterized in that the property information includes the building structure, identification information related to the components used, materials, specifications, usage locations, usage methods, manufacturing locations, or information related to accessories.

[0013] Furthermore, the derivation system is characterized in that the specifications relating to the components used include information indicating the slope of the roof, the roofing material, or the shape of the composite roof.

[0014] Furthermore, the derivation device disclosed in this application is a derivation device for deriving a transport plan for transporting materials necessary for the construction of a building, comprising: an acquisition means for acquiring target property information, which is property information relating to a building that is the subject of the derivation of the transport plan; and a means for accessing a trained model that stores the relationship between actual property information, which is property information of a completed building, and transport results by the transport means, wherein the trained model is generated by machine learning using training data in which selected elements, which are selected in advance from a plurality of elements included in the actual property information based on their relationship with transport results, are input data, and the corresponding transport results are used as labels, and the device comprises a plan derivation means that, based on the target property information acquired by the acquisition means, refers to the trained model and derives a transport plan for transporting materials necessary for the construction of a building indicated in the target property information.

[0015] Furthermore, the derivation program disclosed herein is a derivation program that derives a transport plan for transporting transport means for transporting materials necessary for the construction of a building to a computer, wherein the computer performs an acquisition step of acquiring target property information, which is property information relating to a building that is the subject of the derivation of the transport plan, and a step of accessing a trained model that stores the relationship between actual property information, which is property information of a completed building, and transport results by the transport means, wherein the trained model is generated by machine learning using training data in which selected elements that have been selected in advance from a plurality of elements included in the actual property information based on their relationship with transport results, and the corresponding transport results are used as labels, and the program derives a transport plan for transporting transport means for transporting materials necessary for the construction of a building indicated in the target property information by referring to the trained model based on the target property information acquired in the acquisition step. [Effects of the Invention]

[0016] The derivation system disclosed in this application refers to a trained model generated by machine learning using selected elements chosen in advance based on their relationship with actual transport records, and derives a transport plan for transporting materials based on property information. As a result, the derivation system disclosed in this application is expected to be able to handle long-term predictions and is expected to have excellent effects such as improving the accuracy of predictions regarding the use of transport means. [Brief explanation of the drawing]

[0017] [Figure 1] This is an explanatory diagram conceptually illustrating an example of an embodiment of the derivation system disclosed in this application. [Figure 2] This is a block diagram conceptually illustrating examples of the configurations of various devices used in the derivation system disclosed in this application. [Figure 3] This is a block diagram conceptually illustrating examples of the configurations of various devices used in the derivation system disclosed in this application. [Figure 4] This flowchart shows an example of the learning process of the derivation device used in the derivation system disclosed in this application. [Figure 5]It is a flowchart showing an example of a plan derivation process used in the derivation system disclosed in the present application. [Figure 6] It is an explanatory diagram showing an example of the output result of the conveyance plan in the derivation system disclosed in the present application.

Mode for Carrying Out the Invention

[0018] Hereinafter, embodiments of the present invention will be described in detail. Note that the following embodiments are examples embodying the present invention and do not have the character of limiting the technical scope of the present invention.

[0019] <Application Example> When constructing a building, the derivation system disclosed in the present application derives, for example, a conveyance plan for a conveyance means for conveying necessary materials from a factory. Hereinafter, an embodiment in which the derivation system disclosed in the present application is applied to the formulation of a conveyance plan for a truck will be exemplified and described with reference to the drawings.

[0020] <System Configuration> FIG. 1 is an explanatory diagram conceptually showing an example of an embodiment of the derivation system disclosed in the present application. The derivation system includes a derivation device 1 that executes main processes in the system. The derivation device 1 is connected to a communication network NW such as an in-house LAN (Local Area Network), WAN (Wide Area Network), dedicated communication network, or the Internet. A management device 2 used by a data management person and an order placement device 3 used by an order placement person are connected to the communication network NW. <s

[0021] Furthermore, various computers, such as a database server computer 4 (hereinafter referred to as database server 4) and an inference server computer 5 (hereinafter referred to as inference server 5), are connected to the communication network NW. The derivation device 1 can access these database servers 4 and inference servers 5 and utilize the information stored therein. Figure 1 shows an example of a system in which the derivation device 1 accesses devices such as database servers 4 and inference servers 5 via the communication network NW, but database servers 4 and inference servers 5 may also be directly connected to the derivation device 1 without going through the communication network NW. Moreover, it is also possible to store some or all of the database servers 4 and inference servers 5 within the derivation device 1 and configure it to be accessible as needed.

[0022] <Hardware configuration of each device> Next, we will describe examples of the configurations of various devices used in the derivation system. Figures 2 and 3 are block diagrams conceptually showing examples of the configurations of various devices used in the derivation system disclosed in this application. The derivation device 1 is configured using a computer such as a server computer capable of communication. The derivation device 1 includes various components such as a control unit 10, a storage unit 11, and a communication unit 12.

[0023] The control unit 10 is a processor such as a CPU (Central Processing Unit) that controls the entire device and is equipped with various circuits such as information processing circuits, timing circuits, and register circuits.

[0024] The storage unit 11 is a storage unit composed of non-volatile memory such as hard disks, SSDs (Solid State Drives), RAID (Redundant Arrays of Inexpensive Disks), and flash memory, and volatile memory such as various types of RAM (Random Access Memory). The storage unit 11 stores computer programs (hereinafter referred to as "programs") such as basic programs (OS: Operating System) and application programs that run on the basic program. As application programs, various programs such as the derivation program 110 for realizing the derivation device 1 disclosed in this application are stored. The derivation program 110 includes programs that perform various processes such as learning processes and plan derivation processes.

[0025] The communication unit 12 is a communication device such as a LAN adapter, which connects to the communication network NW and accesses various devices, the database server 4, the inference server 5, and other computers.

[0026] A computer equipped with the various configurations exemplified above operates as a derivation device 1 by reading various programs, such as the derivation program 110, stored in the memory unit 11 under the control of the control unit 10, and executing various procedures contained in the read programs. In the following description, an example of a configuration in which the derivation device 1 is set up as a single device will be used, but the derivation device 1 disclosed in this application can also be set up as a collection of multiple devices. For example, the derivation device 1 can be set up with a derivation device 1 that performs learning processing and a derivation device 1 that performs planning derivation processing as separate devices. Furthermore, the derivation device 1 disclosed in this application can also be used in conjunction with other devices such as a management device 2 and an ordering device 3.

[0027] Management device 2 is a computer-based device, such as a client computer, used by data management personnel. Management device 2 is used for tasks and / or processing such as inputting various data into various databases such as the residence information database 40 and the transport performance database 41 provided by the database server 4, and constructing trained models 50. Management device 2 is equipped with various components such as a control unit 20, a storage unit 21, an input unit 22, a display unit 23, and a communication unit 24.

[0028] The input unit 22 is a device used for inputting operations such as a keyboard or mouse. The display unit 23 is a display device such as a liquid crystal display.

[0029] The ordering device 3 is a computer-based device, such as a client computer, used by ordering personnel. The ordering device 3 is used for tasks related to planning the transportation of materials using transport means such as trucks, using the building information database 40 and the trained model 50. The ordering device 3 is equipped with various components such as a control unit 30, a storage unit 31, an input unit 32, a display unit 33, and a communication unit 34.

[0030] Next, we will describe the various databases provided by the database server 4. The property information database 40 stores various property information associated with a property ID that identifies the property. The property information includes various elements of information such as the building structure, construction method, building materials and identification information related to the materials used, materials, specifications, usage locations, usage methods, manufacturing location, and accessories. The property information is stored as text information in a database stored in tabular format, as drawing information showing building plans and information that can be read from the drawing information, and as aggregated information of these pieces of information.

[0031] Building structure refers to information about the building's structure, such as the number of units, number of floors, style, and whether it is steel-framed or wooden. Building construction method refers to information about the construction method, such as construction classification and construction method. Regarding materials used in the building and information about those materials, identification information refers to information about items such as product codes that identify the materials used, for example, information that identifies individual products. Even for materials of the same type, unique numbers, codes, symbols, etc., are assigned to each element such as specifications, construction direction, and structure. Material refers to the material of the materials used, for example, concrete, gypsum, wood, etc. Specifications refer to the specifications of the materials used, for example, information about items such as roof slope, roofing material, and composite roof shape. Location of use refers to information about items such as roof, second floor, entrance, etc. Method of use refers to information about items for special uses. Place of manufacture refers to information about items such as the factory, workshop, or subcontractor where the goods were manufactured. Accessories refer to information about items such as solar panels and distribution boards.

[0032] The transport performance database 41 is a database that stores transport performance data. The transport performance database 41 stores various transport performance information associated with the item ID. The transport performance information includes information on various items such as the number of transports, transport intervals, and transport capacity. The number of transports is the number of times, for example, that components were transported to the item by truck. The transport interval is information indicating the number of days between each transport. The transport capacity is the capacity required for transport, and in the case of transport by truck, it indicates the number of trucks and the load capacity.

[0033] Next, we will explain the pre-trained model 50 provided by the inference server 5. The pre-trained model 50 consists of multiple models, such as the transport count model 50a, the transport interval model 50b, and the transport capacity model 50c. The transport count model 50a is a model generated by machine learning using the transport count as training data. The transport interval model 50b is a model generated by machine learning using the transport interval as training data. The transport capacity model 50c is a model generated by machine learning using the transport capacity as training data.

[0034] <Software processing for the device> Next, the processing of various devices used in the derivation system disclosed in this application will be explained. First, the learning process will be explained. The learning process is the process of generating a trained model 50 based on the contents of various databases stored in the database server 4. In the derivation system disclosed in this application, the learning process is executed by the derivation device 1 when it receives a command from the management device 2 operated by the data management person, or when it reaches a predetermined time, or when other start conditions are met.

[0035] Figure 4 is a flowchart showing an example of the learning process of the derivation device 1 used in the derivation system disclosed in this application. The derivation device 1 performs the learning process under the control of the control unit 10 which executes the derivation program 110. The control unit 10 of the derivation device 1 acquires property information (actual property information) related to buildings stored in the property information database 40 (S101). In step S101, for example, property information of buildings that have already been completed and whose transport records are stored in the transport record database 41 is extracted as actual property information, which is property information of completed buildings. The extracted actual property information includes information in text format. Furthermore, the control unit 10 acquires actual transport information stored in the transport record database 41 (S102). In step S102, actual transport information indicating the number of transports, transport intervals, and transport capacity is extracted.

[0036] The control unit 10 associates the acquired actual property information and actual transport information based on the property ID and derives the contribution rate of each element of the actual property information to the actual transport information (S103). The derivation of the contribution rate in step S103 is performed, for example, by regression analysis such as multiple regression analysis with actual property information as the explanatory variable and actual transport information as the dependent variable.

[0037] Furthermore, the control unit 10 selects elements of the actual literature information that have a high derived contribution rate, for example, elements of the actual property information included in the top 60%, as selection elements (S104), and stores the selection elements in the property information database 40 (S105). The processing in steps S103 to S105 is performed for each of the actual transport information's transport count, transport interval, and transport capacity, and selection elements that will serve as explanatory variables are selected for each of the transport count, transport interval, and transport capacity. For example, for actual property information related to the transport count, approximately 50 to 100 items such as product code, roof slope, number of units, number of floors, total floor area, B-type foundation specifications, B-type structural beam XX factory, sanitary ware, and distribution board are selected as selection elements. Also, for example, for actual property information related to the transport interval, approximately 50 to 100 items such as product code, construction classification, roof slope, roofing material, number of units, number of floors, load-bearing wall, B-type foundation specifications, B-type structural beam XX factory, and Japanese-style room ceiling are selected as selection elements. Furthermore, for example, regarding actual project information related to transport capacity, approximately 50 to 100 items can be selected as selection factors, such as product code, roofing material, roof composite, number of units, total floor area, underfloor inspection hatch, Type B foundation specifications, Type B structural beam △△ factory, unit bath, wall and ceiling-mounted ventilation fan, etc.

[0038] The control unit 10 obtains selection elements from the property information database 40 (S106) and actual transport information from the transport performance database 41 (S107). Furthermore, the control unit 10 associates the obtained selection elements and actual transport information based on the property ID (S108), and generates training data using the selection elements of the actual property information as input data and the corresponding actual transport information as labels (S109). Then, the control unit 10 generates a trained model 50 by machine learning using the training data (S110), and stores the generated trained model 50 in the inference server 5 (S111). The processing in steps S106 to S111 generates trained models 50 for each of the selection elements of transport count, transport interval, and transport capacity as a transport count model 50a, a transport interval model 50b, and a transport capacity model 50c.

[0039] As described above, the derivation system disclosed in this application performs a learning process to generate trained models 50, such as a transport count model 50a used for AI prediction of the transport count, a transport interval model 50b used for AI prediction of the transport interval, and a transport capacity model 50c used for AI prediction of the transport capacity.

[0040] Next, the transport prediction process will be described. The plan derivation process is the process of deriving a transport plan using a trained model 50 stored in the inference server 5. In the derivation system disclosed herein, the plan derivation process is executed by the derivation device 1 when it receives an order from the ordering device 3 operated by the ordering person, or when it meets start conditions such as reaching a predetermined time.

[0041] Figure 5 is a flowchart showing an example of the plan derivation process used in the derivation system disclosed in this application. The derivation device 1 executes the plan derivation process under the control of the control unit 10 which executes the derivation program 110. The control unit 10 of the derivation device 1 obtains from the building information database 40 the target building information (target building information) that is used for inferring the number of transports from the building information information (S201). In step S201, the building information that is the target building information that is the target building information is extracted as target building information. The target building information extracted for inferring the number of transports is, for example, the target building information that corresponds to the items of the actual building information that were selected as selection elements related to the number of transports in the learning process. The extracted target building information includes information such as information in text format.

[0042] The control unit 10 standardizes the acquired target property information for inference using the trained model 50 (S202). The control unit 10 performs a first inference process to derive the predicted number of transports based on the acquired target property information, by referring to the transport count model 50a of the trained model 50 (S203). The control unit 10 performs a standardization return on the result of the predicted number of transports obtained in the first inference process (S204), and stores the information after the standardization return in the storage unit 11 as the result of the transport count (S205). In this way, the number of transports for the building that is the target of the transport plan is derived.

[0043] The control unit 10 obtains target property information used for inferring the transport interval from the property information database 40 (S206), and further obtains the result of the transport count stored in step S205 from the storage unit 11 (S207). In step S206, the target property information extracted for inferring the transport count is, for example, the target property information corresponding to the items of the actual property information selected as selection elements related to the transport interval in the learning process. The extracted target property information includes information such as text format.

[0044] The control unit 10 standardizes the acquired target property information for inference using the trained model 50 (S208). The control unit 10 performs a second inference process to derive a predicted transport interval based on the acquired target property information and the derived result of the acquired transport count, by referring to the transport interval model 50b of the trained model 50 (S209). The control unit 10 performs a standardization return on the derived result of the predicted transport interval obtained in the second inference process (S210), and stores the information after the standardization return in the storage unit 11 as the transport interval derivation result (S211). In this way, the transport interval for the building that is the target of the transport plan derivation is derived.

[0045] The control unit 10 obtains target property information from the property information database 40 that will be used for inferring transport capacity (S212), and further obtains the result of the transport count stored in step S205 from the storage unit 11 (S213). In step S212, the target property information extracted for inferring transport capacity is, for example, the target property information corresponding to the items of the actual property information selected as selection elements related to transport capacity in the learning process. The extracted target property information includes information such as information in text format.

[0046] The control unit 10 standardizes the acquired target property information for inference using the trained model 50 (S214). The control unit 10 performs a third inference process to derive the predicted transport capacity based on the acquired target property information and the derived result of the acquired transport count, by referring to the transport capacity model 50c of the trained model 50 (S215). The control unit 10 performs a re-standardization of the derived result of the predicted transport capacity obtained in the third inference process (S216), and stores the re-standardized information in the storage unit 11 as the transport capacity derivation result (S217). In this way, the derived capacity for the building that is the target of transport plan derivation is derived.

[0047] The control unit 10 derives a transport plan (S218) based on the transport count, transport interval, and transport capacity stored in the storage unit 11, through processing such as BI tool (business intelligence tools) processing. The derivation of the transport plan in step S218 is a process that derives, for example, the transport date, load capacity, and number of trucks to be used for transport, based on the transport count, transport interval, and transport capacity. In particular, by deriving the transport plan using a BI tool, the transport plan is derived as information in a format that is easy for ordering personnel to understand and easy to use for secondary purposes such as editing and analysis.

[0048] The control unit 10 then outputs the derived transport plan (S219). The output of the transport plan in step S219 is processed as transmission to the ordering device 3, uploading to a communication network NW such as the company's LAN, and storage in the storage unit 11 and a predetermined database.

[0049] Figure 6 is an explanatory diagram showing an example of the output results of the transport plan in the derivation system disclosed in this application. Figure 6 shows an example of the output plan displayed on the display unit of the ordering device 3 as an example of the output results of the transport plan. Figure 6 shows the number of trucks used for transport and the types of materials to be transported, graphed on a daily basis.

[0050] As described above, the derivation system disclosed in this application, based on the target property information, refers to the trained model 50 and performs a plan derivation process to derive a transport plan for transporting transport means that transport materials necessary for the construction of the building indicated in the target property information.

[0051] As detailed above, the derivation system disclosed in this application derives a transport plan by referring to a trained model 50 generated by machine learning using selected elements chosen in advance based on their relationship with transport performance. As a result, the derivation system disclosed in this application can handle long-term predictions and is expected to improve the prediction accuracy related to the prediction of the use of transport means, thus providing excellent effects.

[0052] The present invention is not limited to the embodiments described above and can be implemented in various other forms. Therefore, these embodiments are merely illustrative in all respects and should not be interpreted restrictively. The scope of the present invention is defined by the claims and is not restricted in any way by the text of the specification. Furthermore, any modifications or changes within the equivalent scope of the claims are all within the scope of the present invention.

[0053] The following additional information is disclosed regarding the technical details described in the embodiments described above.

[0054] (Note 1) A derivation system using a derivation device for deriving a transport plan for transporting materials necessary for building construction, A means for acquiring target property information, which is property information relating to the building that is the subject of the derivation of the transportation plan, A trained model that stores the relationship between actual property information (information on completed buildings) and actual transportation data by transportation methods. Equipped with, The aforementioned trained model is generated by machine learning using selected elements, which are pre-selected from multiple elements included in the actual property information based on their relationship with actual transportation records, as input data, and the corresponding transportation records as labels. The derivation device is, Based on the target property information acquired by the acquisition means, the plan derivation means refers to the trained model and derives a transport plan for transporting materials necessary for constructing the building indicated in the target property information. A derivation system characterized by the following:

[0055] (Note 2) The derivation system described in Appendix 1, The aforementioned trained model uses multiple elements included in the actual property information as explanatory variables and selects elements as selected elements based on a multiple regression analysis in which the contribution rate of the transportation performance is higher than a predetermined value, with the transportation performance as the dependent variable. A derivation system characterized by the following:

[0056] (Note 3) The derivation system described in Appendix 1 or Appendix 2, The aforementioned trained model is The transport count, transport interval, and transport capacity from the transport record are used as training data to generate the respective data. The aforementioned plan derivation means is, Based on the derived results for the number of transports, transport intervals, and transport capacity, a transport plan is derived. A derivation system characterized by the following:

[0057] (Note 4) The derivation system described in Appendix 1 or Appendix 2, The aforementioned trained model is A transport count model generated using the transport count data as training data, A transport interval model is generated using the transport intervals of actual transport data as training data, A transport capacity model generated using transport capacity from actual transport data as training data and Includes, The aforementioned plan derivation means is The transport count is derived based on the target property information acquired by the acquisition means, by referring to the transport count model. The transport interval is derived based on the target object information acquired by the acquisition means and the number of transports derived by the acquisition means, referring to the transport interval model. The transport capacity is derived based on the target object information acquired by the acquisition means and the number of transports derived, with reference to the transport capacity model. Based on the derived transport interval and transport capacity, a transport plan is derived. A derivation system characterized by the following:

[0058] (Note 5) A derivation system described in any one of Appendix 1 to Appendix 3, The aforementioned property information includes the building structure, as well as identification information, materials, specifications, usage locations, usage methods, manufacturing locations, and information relating to the components used, or accessories. A derivation system characterized by the following:

[0059] (Note 6) The derivation system described in Appendix 5, The specifications for the components used include information indicating the slope of the roof, the roofing material, or the shape of the composite roof. A derivation system characterized by the following:

[0060] (Note 7) A derivation device for deriving a transport plan for transporting materials necessary for building construction, A means for acquiring target property information, which is property information relating to the building that is the subject of the derivation of the transportation plan, A means to access a trained model that stores the relationship between actual property information, which is property information of completed buildings, and actual transportation records by transportation means. Equipped with, The aforementioned trained model is generated by machine learning using selected elements, which are pre-selected from multiple elements included in the actual property information based on their relationship with actual transportation records, as input data, and the corresponding transportation records as labels. Based on the target property information acquired by the acquisition means, the plan derivation means refers to the trained model and derives a transport plan for transporting materials necessary for constructing the building indicated in the target property information. An extraction device characterized by the following features.

[0061] (Note 8) A derivation program for a computer that derives a transport plan for transporting materials necessary for the construction of a building, On the computer, The acquisition step involves obtaining target property information, which is property information related to the building that is the subject of the transportation plan derivation, and The steps include accessing a trained model that stores the relationship between actual property information (information on completed buildings) and actual transportation data by transportation methods, and It is configured to execute, The aforementioned trained model is generated by machine learning using selected elements, which are pre-selected from multiple elements included in the actual property information based on their relationship with actual transportation records, as input data, and the corresponding transportation records as labels. Based on the target property information acquired in the acquisition step, the trained model is referenced to derive a transport plan for transporting materials necessary for the construction of the building indicated in the target property information. A derivation program characterized by the following. [Explanation of symbols]

[0062] 1 Derivation device 10 Control Unit 11 Storage section 110 Derivation Program 12 Communications Department 2 Management device 20 Control Unit 21 Memory section 22 Input section 23 Display section 24 Communications Department 3. Equipment for ordering 30 Control Unit 31 Storage section 32 Input section 33 Display section 34 Communications Department 4. Database Server Computer 40 House Information Database 41. Transportation Performance Database 5. Inference Server Computer 50 pre-trained models 50a transport count model 50b Conveying interval model 50cc transport capacity model Network

Claims

1. A derivation system using a derivation device for deriving a transport plan for transporting materials necessary for building construction, A means for acquiring target property information, which is property information relating to the building that is the subject of the derivation of the transportation plan, A trained model that stores the relationship between actual property information (information on completed buildings) and actual transportation data by transportation methods. Equipped with, The aforementioned trained model is generated by machine learning using selected elements, which are pre-selected from multiple elements included in the actual property information based on their relationship with actual transportation records, as input data, and the corresponding transportation records as labels. The derivation device is, Based on the target property information acquired by the acquisition means, the plan derivation means refers to the trained model and derives a transport plan for transporting materials necessary for constructing the building indicated in the target property information. A derivation system characterized by the following:

2. The derivation system according to claim 1, The aforementioned trained model uses multiple elements included in the actual property information as explanatory variables and selects elements as selected elements based on a multiple regression analysis in which the contribution rate of the transportation record is higher than a predetermined value, with the transportation record as the dependent variable. A derivation system characterized by the following:

3. A derivation system according to claim 1 or claim 2, The aforementioned trained model is The transport count, transport interval, and transport capacity from the transport record are used as training data to generate the respective data. The aforementioned plan derivation means is Based on the derived results for the number of transports, transport intervals, and transport capacity, a transport plan is derived. A derivation system characterized by the following:

4. A derivation system according to claim 1 or claim 2, The aforementioned trained model is A transport count model generated using the transport count data as training data, A transport interval model is generated using the transport intervals of actual transport data as training data, A transport capacity model generated using transport capacity from actual transport data as training data and Includes, The aforementioned plan derivation means is The transport count is derived based on the target property information acquired by the acquisition means, by referring to the transport count model. The transport interval is derived based on the target object information acquired by the acquisition means and the number of transports derived by the acquisition means, referring to the transport interval model. The transport capacity is derived based on the target object information acquired by the acquisition means and the number of transports derived, with reference to the transport capacity model. Based on the derived transport interval and transport capacity, a transport plan is derived. A derivation system characterized by the following:

5. A derivation system according to claim 1 or claim 2, The aforementioned property information includes the building structure, as well as identification information, materials, specifications, usage locations, usage methods, manufacturing locations, and information relating to the components used, or accessories. A derivation system characterized by the following:

6. The derivation system according to claim 5, The specifications for the components used include information indicating the slope of the roof, the roofing material, or the shape of the composite roof. A derivation system characterized by the following:

7. A derivation device for deriving a transport plan for transporting materials necessary for building construction, A means for acquiring target property information, which is property information relating to the building that is the subject of the derivation of the transportation plan, A means to access a trained model that stores the relationship between actual property information, which is property information of completed buildings, and actual transportation records by transportation means. Equipped with, The aforementioned trained model is generated by machine learning using selected elements, which are pre-selected from multiple elements included in the actual property information based on their relationship with actual transportation records, as input data, and the corresponding transportation records as labels. Based on the target property information acquired by the acquisition means, the plan derivation means refers to the trained model and derives a transport plan for transporting materials necessary for constructing the building indicated in the target property information. An extraction device characterized by the following features.

8. A derivation program for a computer that derives a transport plan for transporting materials necessary for the construction of a building, On the computer, The acquisition step involves obtaining target property information, which is property information related to the building that is the subject of the transportation plan derivation, and The steps include accessing a trained model that stores the relationship between actual property information (information on completed buildings) and actual transportation data by transportation methods, and It is configured to execute, The aforementioned trained model is generated by machine learning using selected elements, which are pre-selected from multiple elements included in the actual property information based on their relationship with actual transportation records, as input data, and the corresponding transportation records as labels. Based on the target property information acquired in the acquisition step, the trained model is referenced to derive a transport plan for transporting materials necessary for the construction of the building indicated in the target property information. A derivation program characterized by the following.

Images