JP2021068010A - Plan generation system construction support device, and plan generation system construction support method - Google Patents

Abstract

To provide a plan generation system construction support device capable of solving a disadvantage that a lot of man-hours is required to implement a system in each factory when constructing a plan generation system in multiple factories.SOLUTION: Before implementing a system at a factory that will build the plan generation system from now on, a conventional mathematical model that represents the characteristics of plans that have been generated and operated and a mathematical model corresponding to the processing of the source code which is created when the plan generation system was already built-in another factory are compared, and when it is determined that there exists a similar mathematical model, the corresponding past source code is reused.SELECTED DRAWING: Figure 3

Description

The present invention relates to a plan generation system construction support device and a plan generation system construction support method.

In the industrial world, it is necessary to generate various plans such as product production plans, delivery plans, and personnel plans at work sites, and there is a system that automatically performs them (hereinafter referred to as a plan generation system).

In providing the plan generation system to the customer, the constraint conditions obtained through the interview with the customer and multiple plan patterns extracted by machine learning from the historical data of the plan previously made by the expert are input. , We provide a system that uses mathematical optimization to generate plans according to the business content of each customer.

In this system development, it is necessary to create a plug-in (program) specific to each customer's case. For example, in order to generate plans at multiple factories of the same company, we decided to build a plan generation system tailored to each factory. Become. However, it takes man-hours to build each of them from scratch, so it is desirable to create one source code for common functions and then reuse it.

In Patent Document 1, from the main keywords commonly included in the source code and specifications created in the past, the past source code and specifications similar to the processing contents of the source code and specifications to be newly implemented from now on. The technology for identifying and reusing it as a template is disclosed.

International Publication No. 2017/199443

However, in order to reuse the existing software in the technique disclosed in Patent Document 1, the past resources must be given a keyword indicating their processing contents in advance.

On the other hand, the development of the above-mentioned plan generation system, which is the target, is implemented flexibly and agilely while listening to the customer's request, and it is not possible to leave a detailed explanation of the function in Japanese for each function of the program. Not. Therefore, the method of selecting the past source code to be used as a template by keywords as in Patent Document 1 cannot be adopted.

The present invention has been made in view of such a background, and is for reducing the source code development manpower by reusing the source code of the same processing contents with respect to the source code of the plan generation system applied at a plurality of places. An object of the present invention is to provide a plan generation system construction support device and a plan generation system construction support method.

A preferred example of the plan generation system construction support device of the present invention is a device that outputs reusable source code among the source codes created in the past when constructing a plan generation system of a client company, and is a device constructed in the past. In the generation system, a mathematical model that expresses the rules and patterns based on the actual data analysis related to the plan generation of the created client company, a correspondence table storage unit that stores the created source code in association with each other, and a second newly constructed second Regarding the plan generation system of the client company, a mathematical model is created from the actual data related to the plan generation of the second client company by using the analysis program used to create the mathematical model related to the plan generation system constructed in the past. A statistical analysis unit that compares and judges the significant difference between the model output unit to be used, the mathematical model related to the newly constructed plan generation system of the second client company, and the mathematical model related to the plan generation system constructed in the past. If the significant difference between the mathematical models is within a predetermined threshold, it is determined that there is no significant difference, and the program that implements the processing of the mathematical model related to the plan generation system of the second client company to be newly constructed is described above. It is configured with a source code output unit that presents the reuse of the source code that implements the mathematical model of the plan generation system constructed in the past searched by the corresponding mathematical model from the correspondence table.

Further, as another feature of the present invention, in the plan generation system construction support device, the constraint conditions required for the plan generation system of the client company constructed in the past and each constraint condition in the correspondence list with the client company are included. An analysis program for creating a mathematical model based on the above is registered, and the model output unit creates the mathematical model from the actual data related to the plan generation of the second client company to be newly constructed. Read and use the analysis program corresponding to the constraint condition with the same function form from the correspondence list with the company.

Further, a preferable example of the plan generation system construction support method of the present invention is a method of outputting a reusable source code from the source code created in the past when constructing the plan generation system of the client company, which is constructed in the past. In the created plan generation system, a process of creating a correspondence table that stores the created source code in association with the mathematical model that expresses the rules and patterns based on the actual data analysis related to the created plan generation of the client company, and newly constructed Regarding the plan generation system of the second client company, the analysis program used to create the mathematical model related to the plan generation system constructed in the past is used to obtain the actual data related to the plan generation of the second client company. Both the process of creating a mathematical model and the process of comparing and judging the significant difference between the mathematical model related to the plan generation system of the second client company to be newly constructed and the mathematical model related to the plan generation system constructed in the past. If the significant difference between the mathematical models is within a predetermined threshold, it is determined that there is no significant difference, and the program that implements the processing of the mathematical model related to the plan generation system of the second client company to be newly constructed is described above. It has a process of presenting the reuse of the source code that implements the mathematical model of the plan generation system constructed in the past, which is searched by the corresponding mathematical model from the correspondence table.

According to the present invention, when constructing a plan generation system in a plurality of factories, it can be performed with a low man-hour.

It is a functional block diagram of the plan generation system construction support device. It is a figure which shows the plan generation system construction procedure. It is a functional block diagram explaining the relationship between each functional part and each storage part in a plan generation system construction support apparatus. This is an example of the constraint condition table of the i-th factory, which is the input for building the plan generation system of the i-th factory. This is an example of a list of factories that comply with constraints. It is a flowchart explaining the process of the constraint condition comparison part and the constraint condition correspondence factory list update part. This is an example of the constraint condition compatible factory list updated by the constraint condition compatible factory list update section. FIG. 8A is an example of a constraint condition violation model when the constraint condition is a + b <α, and FIG. 8B is a diagram illustrating an example of a method of obtaining a significant difference between the two planning pattern models. is there. This is an example of a known plan pattern model table. This is an example of the planning pattern model-source code correspondence table. Source code format This is an example of the source code format output by the output unit.

Hereinafter, embodiments will be described in detail with reference to the drawings.

FIG. 1 shows a functional configuration diagram of the plan generation system construction support device (1) shown as an embodiment. As the plan generation system construction support device (1), a computer including a memory (10), an auxiliary storage device (11), a processor (12), an input device (13), and a display device (14) is used. .. This computer may be a general-purpose computer. Each component is communicably connected to each other via a communication means such as a bus (not shown).

The memory (10) includes a ROM which is a non-volatile storage element and a RAM which is a volatile storage element. The auxiliary storage device (11) is a large-capacity and non-volatile storage device such as a magnetic storage device (HDD) or a flash memory (SSD), and is used when executing a program executed by the processor (12) and a program. Store data.

The processor (12) loads the program (not shown) stored in the auxiliary storage device (11) into the memory (10) and executes it to correspond to the constraint condition comparison unit (101) and the constraint condition. Factory list update unit (102), plan pattern model output unit (103), plan pattern model statistical analysis unit (104), known plan pattern model table update unit (105), and source code format output unit (106). ) Is realized.

Further, as shown in FIG. 1, the auxiliary storage device (11) includes a constraint condition-compliant factory list storage unit (111), a verification target plan pattern storage unit (112), and a known plan pattern model table storage unit (113). And a plan pattern model-source code correspondence table storage unit (114).

FIG. 2 describes a procedure for constructing a plan generation system in a certain factory before applying the plan generation system construction support device (1) of this embodiment. Here, "one factory" is an example of one of a plurality of factories in a certain company, but the example is not limited to this, and multiple factories / business establishments of unrelated companies, etc. It may be the construction of a plan generation system in.

In step S201, as a preliminary step of constructing the plan generation system, the constraint conditions of the plan to be output from the system customer are heard. Since this is an example of implementation, it is not limited to constraints, but it is to hear the information necessary for plan generation.

Next, in step S202, the plans created and used at the factory in the past are analyzed, and the level at which the constraints have been achieved or whether they have been violated is visualized and verified on a graph or the like.

Further, in step S203, the engineer creates an analysis program corresponding to the constraint condition based on the information visualized in S202, and constructs a constraint condition violation model (mathematical model).

Finally, in step S204, a project-specific plug-in is created based on the constraint condition violation model constructed in S203, and the plan generation system is implemented.

The plan generation system construction support device (1) of this embodiment simplifies the procedure of the plan generation system construction described in FIG. 2, and is similar to the violation model showing the constraint conditions in the past plan generation system construction. When implementing a plan generation system for a new customer having a constraint condition and its violation model, the source code related to the past plan generation system is reused to reduce the number of steps for constructing the plan generation system.

FIG. 3 is a functional block diagram illustrating the relationship between each functional unit and each storage unit in the plan generation system construction support device (1) of this embodiment. Hereinafter, the description will be given according to the order of processing according to the arrangement of the functional parts in FIG.

In 201 of FIG. 3, a plurality of constraint conditions were extracted as a result of conducting the hearing shown in S201 of FIG. 2 when constructing the plan generation system of a customer's i-factory, and the plan generation system construction support device (1) was used. It is shown to input from the input device (13).

FIG. 4 shows an example of the constraints required for the plan generation in the i-factory extracted by the hearing.
For example, the constraint condition of # 001 is that the number of days from the order of the product to the end of production is a and the number of days that the product is kept in inventory is b as the constraint condition when constructing the system for generating the production plan of the i factory. Then, the content that the total number of days must be less than 5 days is shown by the function: a + b <5.
A, b, d, and e in the constraint condition function of FIG. 4 are defined variables, respectively.

Here, in the following, an embodiment in a situation where the construction of the plan generation system from the first factory to the (i-1) th factory of a certain company has been completed and the construction of the plan generation system of the i-th factory is currently being carried out. I recommend the explanation of.

In FIG. 5, the construction of the plan generation system from the first factory to the (i-1) th factory (hereinafter referred to as i-1 factory) stored in the constraint condition-corresponding factory list storage unit (111) of FIG. 3 is completed. An example of a list of factories that comply with constraints at the stage is shown.

The constraint condition column (301) is a list of all the constraint conditions obtained by hearing up to the i-1 factory. Here, the variables a, b, and d are defined as the number of days from the ordering of the product to the end of production, as described above. In addition, α and β are variables indicating a threshold value.

The analysis program column (302) indicates the ID number of the program that analyzes the violation status of the applicable constraint condition. The analysis program is a program created by a person to calculate a violation model of the constraint condition when the relevant constraint condition first appears in the construction of a plan generation system up to the i-1 factory. Here, the violation model of the constraint condition is given as an example of implementation, but the whole mathematical model that expresses the rules and patterns related to the plan generation is targeted.

The factory No. column (303) shows 1, 2, 3, ... I-1 as an example of the factory ID for which the plan generation system has been constructed so far. For example, column i-1 (304) shows that d ≦ 4100 exists and there is no constraint on a + b in the constraints of the i-1 factory. In this way, the actual values of the threshold values α and β in each factory are shown for each constraint condition.

FIG. 6 shows a flow diagram in which the processing in the constraint condition comparison unit (101) of FIG. 3 and the processing in the constraint condition corresponding factory list update unit (102) are combined.

The constraint condition comparison unit (101) in FIG. 3 inputs the constraint condition (201) of the i factory and the constraint condition compatible factory list (111), and selects the jth constraint condition among the constraint conditions of the i factory. (Constraint (i, j)), whether it is in the constraint condition compatible factory list (111) (whether the constraint condition with the same function form is already registered in the construction of the plan generation system up to i-1 factory Please).

When the constraint condition (i, j) is not in the constraint condition compatible factory list (111) (The form of the function of the constraint condition (i, j) is the form of the function of all the constraint conditions in the constraint condition compatible factory list. (A)), in the constraint condition compatible factory list update unit (102), create a new row (305 in FIG. 7) in the constraint condition compatible factory list (111) and use the constraint condition (i, j). Is newly registered (306 in FIG. 7), a new column of i-factory is created (307 in FIG. 7), and a threshold is registered in the new constraint row of the i-factory column (308 in FIG. 7).

Then, in this case, an analysis program is manually created from S202 of FIG. 2 according to the conventional method. Register the created analysis program in a new line of the constraint condition compatible factory list (111), and move to the planning pattern model output unit (103).

When the constraint condition (i, j) is in the constraint condition compatible factory list (111) (the constraint condition (i, j) and one of the constraint conditions in the constraint condition compatible factory list have the same threshold value. If the function form is the same regardless of (b)), in the constraint condition compatible factory list update section (102), create a new column of i factory in the constraint condition compatible factory list (111) and apply the corresponding constraint. Register the threshold in the condition line. After that, the process shifts to the planning pattern model output unit (103).

FIG. 7 shows an example of the constraint condition compatible factory list (111) in which the threshold values are registered for all the constraint conditions related to the i factory after the update process in the constraint condition compatible factory list update unit (102).

In the planning pattern model output unit (103) of FIG. 3, the analysis program (corresponding constraint condition is the first) described in the analysis program column (302) of the constraint condition corresponding factory list (111) regarding the constraint condition (i, j). The constraint condition (i, j) violation model is calculated using (created when it appeared) and stored in the verification target plan pattern storage unit (112).

FIG. 8A shows an example of a constraint condition violation model when the constraint condition is a + b <α. It is assumed that the analysis program inputs the actual data of the i-factory for a predetermined period, and for example, a histogram graph close to the Gaussian distribution as shown in FIG. In this graph, the region 309 above the threshold value α violates the constraint condition. The top x% on the graph violates. This is called a violation model. In actual factory operation, the constraints are 100% observed in normal times, but for example, it is okay to violate the busy time at the end of the month, emergency, etc., for example, overtime work for 5 days at the end of the month. It is generally practiced that operations such as increasing the number may be made. Often it allows a constraint to be violated somewhere and works better with it. Therefore, it is possible to build a plan generation system by incorporating those violation models.

In the technology of constructing the above-mentioned constraint condition violation model (mathematical model) and implementing the plan generation system based on the constructed constraint condition violation model, there is a plan pattern model as a superordinate concept including the constraint condition violation model. For example, the plan pattern model targets all mathematical models that express the rules and patterns related to manufacturing equipment in factories, human relationships, etc. in constructing a plan generation system. In this embodiment, the constraint condition violation model will be described as an example, but the planning pattern model will be used in the name of each functional unit.

FIG. 9 is an example of a known plan pattern model table stored in the known plan pattern model table storage unit (113) of FIG. The known plan pattern model table was created based on the actual data in each factory to which the constraint conditions are applied for each constraint condition (310) that appeared in the construction of the existing plan generation system from 1 to i-1 factories. The function (312) of the constraint condition violation model (311) and the violation status (313) calculated from the violation model are registered together.

In the planning pattern model statistical analysis unit (104) of FIG. 3, the constraint condition (i, j) violation model stored in the verification target planning pattern storage unit (112) is stored in the known planning pattern model table storage unit (113). Analyze and judge whether there is a statistically significant difference from each constraint condition violation model (311) under the same constraint conditions as the constraint conditions (i, j) registered in the known plan pattern model table. To do.

Here, the "significant difference" is, for example, shown in FIG. 8 (B) with the constraint condition (i, j) violation model function (314) and the constraint condition (i, j) of the known plan pattern model table. The function (315) of each constraint violation model under the same constraint is compared in the same coordinate system. Both functions are created by the same analysis program, and the total area of the difference area (316) of the graphs of both functions is the area formed by the graph of the function (314) of the constraint condition (i, j) violation model. If it is within 2.0% (within the threshold value) of (317), it is judged that there is no significant difference. Such a definition is set in advance in the planning pattern model statistical analysis unit (104).

FIG. 10 is an example of the planning pattern model-source code correspondence table stored in the planning pattern model-source code correspondence table storage unit (114) of FIG. The constraint No. (321) in the first column is the ID number of the constraint condition that has appeared so far, and corresponds to the constraint No. (310) in FIG. The known plan pattern model (322) in the second column is the ID number of the constraint condition violation model of the factory to which the constraint condition specified by the constraint number in the first column is applied, and corresponds to 311 in FIG. There is. The source code (323) in the third column represents the ID number of the source code that implements the process that realizes the violation model in the plan generation system.

If the planning pattern model statistical analysis unit (104) in FIG. 3 determines that the same constraint condition as the constraint condition (i, j) violation model is not registered in the known plan pattern model table, the constraint condition (i, j) is not registered. , j) The source code for the plan generation system that realizes the violation model needs to be newly created.

In this case, as shown in step S204 of FIG. 2, a project-specific plug-in is created based on the constraint condition violation model constructed in S203, and the plan generation system is implemented using the mathematical optimization engine. Do.

If the constraint condition (i, j) violation model and the constraint condition (i, j) registered in the known plan pattern model table are the same as the constraint condition (i, j) by the planning pattern model statistical analysis unit (104) in FIG. If it is determined that there is a statistically significant difference from each constraint violation model (311), it is necessary to create a new source code for the plan generation system that realizes the constraint condition (i, j) violation model. ..

In this case, as shown in step S204 of FIG. 2, a project-specific plug-in is created based on the constraint condition violation model constructed in S203, and the plan generation system is implemented using the mathematical optimization engine. Do.

If the constraint condition (i, j) violation model and the constraint condition (i, j) registered in the known plan pattern model table are the same as the constraint condition (i, j) by the planning pattern model statistical analysis unit (104) in FIG. There is no statistically significant difference from any of the constraint condition violation models (311), and the violation status of the corresponding constraint condition violation model (313) and the violation status of the constraint condition (i, j) violation model are predetermined. If it falls within the error range within the threshold, the source code format output unit (106) in FIG. 3 inputs the planning pattern model-source code correspondence table, and there is no significant difference from the constraint condition (i, j) violation model. The source code (323) corresponding to the known plan pattern model (322) is extracted, and the corresponding source code is output to the display device (14) of FIG.

If the constraint condition (i, j) violation model and the constraint condition (i, j) registered in the known plan pattern model table are the same as the constraint condition (i, j) by the planning pattern model statistical analysis function (104) in FIG. There is no statistically significant difference from multiple constraint condition violation models (311), and the violation status of the corresponding constraint condition violation model (313) and the violation status of the constraint condition (i, j) violation model are predetermined thresholds. If the error range is within the range, the above processing is performed on the violation model having the smallest significant difference.

Known after the comparison judgment process between the constraint condition (i, j) violation model and the constraint condition violation model registered in the known plan pattern model table is completed by the planning pattern model statistical analysis unit (104) in FIG. The plan pattern model table update unit (105) stores the constraint condition (i, j) violation model stored in the verification target plan pattern storage unit (112) in the known plan pattern model table storage unit (113). It is newly registered as a constraint condition violation model (311) in the corresponding constraint condition in the known plan pattern model table.

In addition, when the source code of the process that realizes the constraint condition (i, j) violation model is newly created, the known plan pattern model table update unit (105) is stored in the verification target plan pattern storage unit (112). The constraint condition (i, j) violation model is stored in the known plan pattern model table storage unit (113). A new constraint condition registration table is provided in the known plan pattern model table, and the constraint condition violation model in the corresponding constraint condition is provided. Register as (311).

Further, the known plan pattern model table update unit (105) makes the plan pattern model-source code correspondence table storage unit (114) correspond to each item of constraint No. (321) and known plan pattern model (322). , Register the newly created source code (323).

FIG. 11 is a schematic example of the entire source code finally displayed on the display device (14) of FIG.
Both S1 (plan) (331) and S2 (plan) (332) are the constraint condition (i, j) violation models registered in the known plan pattern model table by the planning pattern model statistical analysis unit (104). The constraint condition violation model that has no statistically significant difference and the violation status of the violation model matches is searched, and the source code (323) corresponding to the corresponding constraint condition violation model is output by the source code format output unit (106). It is extracted and displayed on the display device (14) after the function 001 and the function 002.

In S3 (plan) (333), the planning pattern model statistical analysis unit (104) has the same constraint conditions as the constraint condition (i, j) violation model in the known plan pattern model table, and there is a statistically significant difference. It is judged that there is no constraint condition violation model (311), and the developer creates a new source code and inputs it after the function 003.

S4 (plan) (334) is statistically significantly different under the same constraints as the constraint condition (i, j) violation model registered in the known plan pattern model table by the planning pattern model statistical analysis unit (104). Search for the constraint condition violation model (311) in which the violation status of the constraint condition violation model (313) and the violation status of the constraint condition (i, j) violation model are within the error range within a predetermined threshold. Then, the source code format output unit (106) extracts the source code (323) corresponding to the corresponding constraint condition violation model, and displays it on the display device (14) following the function 004. At that time, the displayed source code is the source code that the top 20% violated, so a message such as "Change the violation threshold to the top 15%" is output to the developer. ..

As described above, according to the present embodiment, it is possible to propose a reusable source code created in the past when constructing the plan generation system.

1 Plan generation system construction support device
10 memory
11 Auxiliary storage
12 processors
13 Input device
14 Display device
101 Constraint comparison section
102 Constraint-compliant factory list update department
103 Planning pattern model output section
104 Planning Pattern Model Statistical Analysis Department
105 Known plan pattern model table update section
106 Source code format output section
111 Constraint-compliant factory list storage unit
112 Verification target plan pattern storage unit
113 Known plan pattern model table storage
114 Planning pattern model-source code correspondence table storage
201 i Factory constraints

Claims (8)

It is a device that outputs reusable source code out of the source code created in the past when constructing a plan generation system of a client company.
In the plan generation system built in the past, a mathematical model that shows the rules and patterns based on the actual data analysis related to the plan generation of the created client company, a correspondence table storage unit that stores the created source code in association with each other,
Regarding the plan generation system of the second customer company to be newly constructed, the analysis program used to create the mathematical model related to the plan generation system constructed in the past is used to relate to the plan generation of the second customer company. A model output unit that creates a mathematical model from actual data,
A statistical analysis unit that compares and judges the significant difference between the mathematical model related to the newly constructed plan generation system of the second client company and the mathematical model related to the plan generation system constructed in the past.
If the significant difference between the two mathematical models is within a predetermined threshold, it is determined that there is no significant difference, and as a program that implements the processing of the mathematical model related to the plan generation system of the second client company to be newly constructed. A source code output unit that presents the reuse of the source code that implements the mathematical model of the plan generation system constructed in the past searched by the corresponding mathematical model from the correspondence table, and
A plan generation system construction support device characterized by being equipped with. Register the analysis program that creates the mathematical model based on each constraint condition in the constraint conditions required for the plan generation system of the client company constructed in the past and the correspondence list with the client company.
When the model output unit creates a mathematical model from the actual data related to the plan generation of the second client company to be newly constructed, the constraint condition having the same function form from the correspondence list between the constraint condition and the customer company The plan generation system construction support device according to claim 1, wherein the analysis program corresponding to the above is read out and used. The constraint conditions that are the basis for creating the mathematical model, the functional form of the mathematical model, and the constraints are related to the mathematical model of the plan generation system constructed in the past and the correspondence table that stores the created source code in association with each other. The plan generation system construction support device according to claim 1, wherein the information on the violation status of the conditions is stored. The process by which the statistical analysis unit compares and determines the significant difference between the mathematical model related to the plan generation system of the second client company newly constructed and the mathematical model related to the plan generation system constructed in the past is a bi-mathematical model. The plan generation system construction support device according to claim 3, wherein the graphs of the functions of the above functions are drawn in the same coordinate system, and it is determined that there is no significant difference when the ratio of the mismatched regions of both graphs is equal to or less than a predetermined threshold value. .. In the process in which the statistical analysis unit compares and determines the significant difference between the mathematical model related to the newly constructed plan generation system of the second client company and the mathematical model related to the plan generation system constructed in the past, both mathematical models are used. The graph of the function of is drawn in the same coordinate system, and in addition to the ratio of the mismatched areas of both graphs being less than or equal to the predetermined threshold, the difference in the violation status of both mathematical models is within the predetermined threshold. The plan generation system construction support device according to claim 3, wherein the device is determined to be none. From the correspondence table, the source code output unit implements the processing of the mathematical model related to the plan generation system of the second client company to be newly constructed according to the determination that there is no significant difference between the two mathematical models. The plan generation according to claim 5, wherein when presenting reuse of the source code, a message of an instruction to change the description of the source code is displayed based on the difference in the violation status of the two mathematical models. System construction support device. It is a method to output reusable source code out of the source code created in the past when constructing a plan generation system of a client company.
In the plan generation system built in the past, the process of creating a correspondence table that stores the created source code in association with the mathematical model that shows the rules and patterns based on the actual data analysis related to the plan generation of the created client company.
Regarding the plan generation system of the second customer company to be newly constructed, the analysis program used to create the mathematical model related to the plan generation system constructed in the past is used to relate to the plan generation of the second customer company. The process of creating a mathematical model from actual data,
A process of comparing and determining a significant difference between a mathematical model related to a newly constructed plan generation system of a second client company and a mathematical model related to a plan generation system constructed in the past.
If the significant difference between the two mathematical models is within a predetermined threshold, it is determined that there is no significant difference, and as a program that implements the processing of the mathematical model related to the plan generation system of the second client company to be newly constructed. The process of presenting the reuse of the source code that implements the mathematical model of the plan generation system constructed in the past searched by the corresponding mathematical model from the correspondence table, and
A plan generation system construction support method characterized by having. Register the analysis program that creates the mathematical model based on each constraint condition in the constraint conditions required for the plan generation system of the client company constructed in the past and the correspondence list with the client company.
The process of creating the mathematical model has a functional form from the list of correspondence between the constraints and the client company when creating the mathematical model from the actual data related to the plan generation of the second client company to be newly constructed. The plan generation system construction support method according to claim 7, wherein an analysis program corresponding to the same constraint condition is read out and used.

Images