TWI818176B - Planning aids and planning aids - Google Patents

Abstract

The present invention provides a plan-making auxiliary device capable of formulating a plan that adjusts an evaluation index and a relaxing amount of a restriction condition based on input information. The plan formulation auxiliary device includes: a generation unit that generates a knowledge model from input history and plan history, where the input history includes constraints and evaluation indicators in each plan, and the plan history includes relevant relaxation information. Amount of information, the relaxation amount represents the amount of violation of the constraints of each plan, the knowledge model represents the relationship between the relaxation amount of the constraints and the evaluation index; and the output part, which outputs the knowledge model to allow the plan formulation department to Using, the planning department uses the knowledge model to formulate plans.

Description

Planning aids and planning aids

The present invention relates to a plan-making assisting device and a plan-making assisting method, for example, a plan-making assisting device and a plan-making assisting method suitable for assisting in making a plan.

There are many situations where advance planning is important, such as the manufacturing of products and the operation and management of large-scale systems. When formulating a plan, it is necessary to establish constraints that comply with various matters such as time, space, equipment, and resources such as people, and also consider maximizing production volume, maximizing equipment operation rate, and minimizing the number of workers. and other evaluation indicators for projects. When making a plan manually from the beginning, it takes too much time, so computers are often used.

On the other hand, in real environments, constraints are often complicated, and it may be difficult to find a solution that adheres to all constraints even with the use of a computer. Even in this case, the planner of the plan (hereinafter referred to as the planner) can judge how much restriction conditions should be relaxed based on long-term knowledge, and formulate a plan that relaxes several restrictions. However, it is difficult for computers to correctly define the constraints required to formulate plans, and the resulting plans are difficult to satisfy the planners.

Regarding this problem, there is a technology that appropriately learns the relaxation amount of the constraint through a learning unit, and formulates a plan based on the relaxation amount of the same constraint for similar input information.

As this technique, it is proposed to simulate cases using the optimal processing order, accept corrections such as the relaxation amount and priority of the planner's constraints with respect to the displayed simulation results, and learn the contents of the accepted corrections. This is a simulation method that re-sets knowledge parameters such as relaxation amount and priority of restriction conditions and simulates again (see Patent Document 1).

In addition, it is proposed to calculate the planning parameters including the production sequence of each product based on historical information about the production plans of each product that were drawn up in the past, taking into account the relaxation amount of each restriction condition when producing each product, and based on the calculated plan. Draw parameters, enumerate the production sequence of each product, develop several plan candidates for the production plan of each product, evaluate the plurality of plan candidates based on the amount of relaxation with respect to each constraint, and select the best plan candidate from the plurality of plan candidates A production plan formulation method for an excellent production plan, etc. (see Patent Document 2).

[Prior technical literature] [Patent Document]

[Patent Document 1] Japanese Patent Application Publication No. 2005-339402

[Patent Document 2] International Publication No. 2018/220744

In the above-described technology, it is possible to formulate a plan with relaxed restrictions based on the plan maker's correction results or plans drawn up in the past.

However, Patent Document 1 assumes that the planner's correction is accepted. For example, the one-day delivery limit can be relaxed when the order is 100 or more, and the two-day delivery limit can be relaxed when the order is 200 or more. The restriction conditions When the amount of input information (in this case, the number of orders) required for the formulation of each plan is different, the planner needs to make corrections each time, which may prevent effective planning.

In addition, Patent Document 1 and Patent Document 2 assume that a plan is developed to minimize or maximize the evaluation index after determining a reference value, a ratio, and the like of the relaxation amount of the restriction conditions as a fixed value. For example, even if the production capacity limit is relaxed when the profit maximization is small relative to the planned profit of the evaluation index, the profit can be increased. However, this cannot be compared with the method of not relaxing the production capacity limit when the profit exceeds the value of the evaluation index. Corresponds to situations where the amount of relaxation of constraints changes dynamically.

The present invention has been made in consideration of the above points, and an object thereof is to provide a plan preparation assisting device and the like capable of formulating a plan that adjusts an evaluation index and a relaxation amount of a restriction condition based on input information.

In order to solve this problem, the present invention includes: a generation unit that generates a knowledge model from input history and plan history, where the input history includes constraints and evaluation indicators in each plan, and the plan The history includes information on the amount of relaxation, which represents the amount of violation of the aforementioned constraints of each of the aforementioned plans, the knowledge model represents the relationship between the amount of relaxation of the aforementioned constraints and the aforementioned evaluation index; and an output unit outputs the aforementioned The knowledge model is used by the planning department, and the planning department uses the aforementioned knowledge model to formulate plans.

In the above structure, a knowledge model representing the relationship between the evaluation index and the relaxation amount of the restriction condition is generated. Based on the knowledge model, for example, the planning department can grasp the relationship (equilibrium) of how much violation is allowed based on how much value the planner has in the past. Therefore, for example, the planning department can formulate a plan that is close to the balance of past plans based on the input information, that is, a plan that reflects the relationship between the evaluation index and the relaxation amount of the restriction condition.

According to the present invention, plan formulation can be assisted.

100: Planning assistance device

111:Generation Department

112:Output Department

113: Planning Department

[Fig. 1] Fig. 1 is a diagram showing an example of the structure of the planning support device according to the first embodiment.

[Fig. 2] Fig. 2 is a diagram showing an example of the data structure of the input history according to the first embodiment.

[Fig. 3] Fig. 3 is a diagram showing an example of the data structure of the plan history according to the first embodiment.

[Fig. 4] Fig. 4 is a diagram showing an example of the data structure of plan candidates according to the first embodiment.

[Fig. 5] Fig. 5 is a diagram showing an example of the processing sequence of the plan preparation assisting method according to the first embodiment.

[Fig. 6] shows the plan history and plan candidates according to the first embodiment The project information is plotted in an image chart with two evaluation index axes.

[Fig. 7] Fig. 7 is an image diagram of changing the value of a label of a plan candidate in the first embodiment.

[Fig. 8] is an image diagram of changing the value of a label of a plan candidate according to the first embodiment.

[Fig. 9] is a diagram showing an example of the processing sequence of the plan preparation assisting method according to the first embodiment.

[Fig. 10] is an image diagram showing each optimal solution probability regarding the plan information of the first embodiment.

[Fig. 11] Fig. 11 is a diagram showing an example of the knowledge model generation screen according to the first embodiment.

[Fig. 12] Fig. 12 is a diagram showing an example of the plan creation screen according to the first embodiment.

Hereinafter, one embodiment of the present invention will be described in detail with reference to the accompanying drawings. This embodiment relates to a technology for formulating a plan with high accuracy by reflecting a solution preferred by the plan maker in the evaluation index of the plan.

(1) First embodiment ---Device structure---

In FIG. 1 , reference numeral 100 generally indicates the planning assistance device according to the first embodiment.

FIG. 1 is a diagram showing an example of the structure of the planning support device 100.

The plan preparation support device 100 is a computer capable of making effective plans by reflecting tacit knowledge that adjusts evaluation indicators and relaxation amounts of constraints based on input information. As a specific structure for realizing the planning support device 100, a main program, a personal computer, etc. can be assumed.

The plan of this embodiment is, for example, a production plan that assumes that goods (products) produced in a manner that maximizes profits within a predetermined production capacity range are selected through a series of processes using various resources such as production equipment and workers. Painting. Therefore, the planning support device 100 in this case is an adjustment amount that derives evaluation indicators and relaxed amounts of constraints that would have been tacit knowledge in the past from a plurality of production plans prepared in the past for product manufacturing. , and apply it to planned processing devices.

In addition, the plan is not limited to the production plan of the product, but may also be the personnel plan of the employees, the deployment plan of the vehicle, etc.

The hardware included in the planning support device 100 is, for example, the components shown in FIG. 1 . That is, the planning support device 100 includes a central processing device 110, a storage device 120, a memory 130, an input device 140, and an output device 150.

The central processing device 110 is a processor such as a CPU (Central Processing Unit). The central processing device 110 performs comprehensive control of the planning support device 100 itself and performs various judgments, calculations, and control processes. The storage device 120 is composed of a non-volatile memory such as an SSD (Solid State Drive) and a magnetic medium such as a hard disk drive. The storage device 120 stores at least an input history 121 (input history information) and a plan history 122 (plans). historical information information), project candidates 123 (project candidate information), knowledge model 124 (knowledge model information), and program 125.

The memory 130 is composed of volatile storage elements such as RAM (Random Access Memory). The input device 140 is a keyboard, a pointing device, a microphone, or the like. The input device 140 accepts key input, voice input, and the like from the user. The output device 150 is a visual data terminal, a speaker, etc. The output device 150 displays various information, outputs sounds, and the like.

The central processing device 110 reads the program 125 stored in the storage device 120 into the memory 130 and executes it, installs the generation unit 111 that executes the knowledge model generation process of generating the knowledge model 124, and outputs the knowledge model 124 generated by the generation unit 111. Functions such as the output unit 112 and the plan formulation unit 113 that formulate plans based on the knowledge model 124 output from the output unit 112.

In addition, the functions (generation unit 111, output unit 112, plan formulation unit 113, etc.) of the plan-making auxiliary device 100 can, for example, read the program 125 stored in the storage device 120 into the memory 130 through the central processing device 110 and It can be implemented by executing (software), or it can be implemented through hardware such as dedicated circuits, or it can be implemented by combining software and hardware. In addition, part of the functions of the plan-making auxiliary device 100 may be realized by another computer capable of communicating with the plan-making auxiliary device 100 .

In addition to being stored in the storage device 120, the program 125 for installing the functions of the planning support device 100 can also be imported from another device to the storage device 120 via a predetermined medium when necessary during execution, etc. by the planning support device 100. . Specified media means, for example, media that can be installed and detached in planning aids. A storage medium that supports a specified interface of the device 100.

In addition, the input history 121 is a collection of past input information. Input information is a collection of information that includes at least the necessary constraints and evaluation indicators in order to formulate a plan. In addition, the plan history 122 is a collection of plan information indicating plans made by past plan makers. In addition, the plan candidates 123 are a set of plan information output by the plan preparation support device 100 . One piece of plan information is generated from one piece of input information. Therefore, the input information of the input history 121, the plan information of the plan history 122, and the plan information of the plan candidate 123 are in a corresponding relationship.

The plan of this embodiment is a production plan in which products are selected to be produced in a manner that maximizes profits within a predetermined production capacity range as described above. At this time, examples of constraints, evaluation indicators, and variables included in the input information are shown in (Formula 1).

I in (Formula 1) is a set of products that become production candidates when planning is formulated. _wi is the cost to produce the i-th product in the production candidate. _vi is the profit obtained by producing the i-th product in the candidate and selling it. W is the maximum value of production capacity and the total limit value of the production costs of multiple commodities within a plan. x _i is a variable that takes the value of "0" or "1". The value "0" of x _i means that the i-th product in the production candidate will not be produced, and the value "1" of x _i means that the i-th product in the production candidate will be produced. In this embodiment, making a plan means determining the value of x _i . Therefore, there is no value for x _i in the input information, and the value of x _i in the plan information is specified as "0" or "1".

In addition, the restriction condition of (Formula 1) is that the total production cost of the goods produced does not exceed the maximum production capacity, but in addition, it may also have combination conditions of the type of goods, inventory conditions, and delivery conditions. period conditions and many other restrictions. In addition, the evaluation index of (Formula 1) is an index for formulating a plan so as to maximize the total value of the products produced, but in addition, it may also include minimizing the number of types of products, the size of the products, and the production cost. Minimization and other evaluation indicators.

In addition, the knowledge model 124 is information that the generation unit 111 calculates the tacit knowledge of the plan maker from the input history 121, the plan history 122, and the plan candidates 123.

---Function---

Next, the function of the planning support device 100 will be described. The functions described below will be explained as functions installed by the planning support device 100 by executing the program 125, for example.

The planning support device 100 uses the input data shown in the above (Formula 1) to The plan candidate 123 is generated by applying a predetermined algorithm to the information (information read from the input history 121). Past plan information shared with the plan candidate 123 and the input information is extracted from the plan history 122, and the past plans are The drawing information and plan candidates 123 are applied to a prescribed algorithm to generate a knowledge model 124 that adjusts the relaxation amount of the constraint as tacit knowledge of the plan maker. The function of generating the knowledge model 124 in this way depends on the generation unit 111 .

In addition, the plan formulation support device 100 outputs the generated knowledge model 124 so that the plan formulation unit 113 can use it. The output method is not limited to a specific method, and examples thereof include storing in the storage device 120, notifying the planning unit 113, sending to other computers, and displaying on the output device 150. The function of such output depends on the output section 112 .

In addition, the plan formulation auxiliary device 100 receives new input information for formulating a new plan from the planner through the input device 140, and applies the new input information to a prescribed algorithm (for example, using the knowledge model 124 ) to formulate a new plan that reflects the tacit knowledge of the plan maker who adjusts the relaxation amount of the constraint, and outputs it through the output device 150 . The function of formulating such a plan depends on the planning unit 113 .

---Data structure example---

Next, a specific example of data used by the planning support device 100 will be described. First, a specific example of the input history 121 will be described.

FIG. 2 is a diagram showing an example of the data structure of the input history 121. The input history 121 is a collection of input information used for past planning as described above.

Each record of the input history 121 illustrated in FIG. 2 is a record in which the values of the plan number 201, the number of product candidates 202, the maximum production capacity 203, the product number 204, the production cost 205, and the value 206 are associated with each other.

The value of project number 201 is the identification information that uniquely determines the project information. As a record of the same project number 201, the project information of the project number 201 is described. The value of the number of product candidates 202 is the value of the number of product candidates produced in the project number 201, and corresponds to the number of records in the project number 201. The value of the maximum production capacity 203 is the limit value of the total production costs of the plurality of products in the plan number 201. The value of product number 204 is the identification information that uniquely identifies the product in the project number 201. The value of production cost 205 is the cost spent to produce the product. The value of 206 is the profit earned by producing the item and selling it.

In addition, the input history 121 stores the restriction conditions, evaluation indicators, and variables (for example, (Formula 1)) in the project number 201 as described above. For example, it can be formed to represent the restriction conditions, evaluation indicators, etc. The formula is stored in a table structure for each row of the record, and can also be other data structures. In addition, the input history 121 mentioned above may also include, in addition to the information about the product, identification information of the planner, operating status of the production equipment, temperature, humidity, weather and other data as data required for making the plan.

Next, a specific example of the plan history 122 will be described.

FIG. 3 is a diagram showing an example of the data structure of the plan history 122. The plan history 122 is a collection of plan information prepared in the past by the plan maker based on the input information described above.

Each record of the project history 122 illustrated in FIG. 3 is project number 301, Records associated with each value of the number of product candidates 302, maximum production capacity 303, product number 304, production cost 305, value 306, selection mark 307, profit 308, relaxation amount 309, and mark 310. The values of the plan number 301, the number of product candidates 302, the maximum production capacity 303, the product number 304, the production cost 305, and the value 306 are the same as the values of the input history 121 described above.

The value of the selection flag 307 is identification information indicating that the planner chooses to produce or not produce the product. When the value of the selection flag 307 is "1", it means production, and when the value of the selection flag 307 is "0", it means no production. The value of profit 308 is the value of the evaluation index of the plan number 301. The value of the profit 308 is the total value of the value 306 of the products produced among the products of the plan number 301 (the products whose value of the selection flag 307 is "1"). The value of the profit 308 can also be calculated by multiplying the value of the value 306 of the product of the project number 301 and the value of the selection flag 307 so as to be the evaluation index of (Formula 1).

The value of the relaxation amount 309 is the amount that violates the restriction condition of the plan number 301 (the amount that exceeds the value of the maximum production capacity 303). The value of the relaxation amount 309 is the total value of the production cost 305 of the products produced from the products of the plan number 301 (products whose value of the selection flag 307 is "1") minus the maximum production of the plan number 301 The value of ability 303.

The value of the relaxation amount 309 is calculated by multiplying the value of the production cost 305 of the product of the project number 301 and the value of the selection flag 307 so that the value of the relaxation amount 309 is within the constraint condition of (Formula 1). The value of the maximum production capacity 303 of the number 301 can also be calculated. That is, when the value of the relaxation amount 309 is a positive value It means that the restriction condition is violated (the value of the maximum production capacity 303 is exceeded). When the value of the relaxation amount 309 is a negative value, it means that the restriction condition is complied with (the value of the maximum production capacity 303 is not exceeded).

The value of the flag 310 is identification information for distinguishing the plan information from the plan candidate 123 output by the plan preparation support device 100 described below. The value of the flag 310 in the project history 122 is stored with "0".

Next, a specific example of the plan candidate 123 will be described.

FIG. 4 is a diagram showing an example of the data structure of the plan candidate 123. The plan candidates 123 are a collection of plan information output by the plan preparation support device 100 based on the above-mentioned input information as described above. In addition, the plan preparation support device 100 outputs plan information that becomes a plurality of candidates from one input information. Therefore, the plan candidates 123 are an aggregate of aggregates of plan information.

Each record of the plan candidate 123 illustrated in FIG. 4 is the plan number 401, the number of product candidates 402, the maximum production capacity 403, the product number 404, the production cost 405, the value 406, the selection flag 407, the profit 408, and the relaxation amount 409. , records associated with each value of identification 410.

The values of the plan number 401, the number of product candidates 402, the maximum production capacity 403, the product number 404, the production cost 405, and the value 406 are the same as the values of the input history 121 described above.

The value of the selection flag 407 is identification information indicating that the planning assistance device 100 selects to produce or not produce the product. When the value of the selection flag 407 is "1", it means that production is performed, and when the value of the selection flag 407 is "0", it means that production is not performed. The values of Profit 408 and Relaxation 409 are consistent with the above plan history 122 The values of Profit 308 and Relaxation 309 are calculated in the same way. In addition, as described above, the plan preparation support device 100 outputs a plurality of candidate plan information from one input information. Therefore, the value of the flag 410 is identification information for specifying each piece of plan information.

---Example of processing sequence---

Hereinafter, the procedure of the planning assistance method of this embodiment will be described using FIGS. 5 to 12 . Various operations corresponding to the plan-making assisting method described below are realized by the plan-making assisting device 100 executing the program 125 . In addition, the program 125 is composed of codes for performing various operations described below.

FIG. 5 is a diagram illustrating an example of a processing sequence (mainly an example of a flowchart of processing executed by the plan-making assisting device 100) of the plan-making assisting method according to this embodiment. The flowchart schematically illustrates how the generation unit 111 of the plan preparation support device 100 generates the knowledge model 124 that adjusts the relaxation amount of the restriction condition from the input history 121 and the plan history 122 accumulated in the storage device 120 as the implicit knowledge of the plan maker. The knowledge output unit 112 stores the knowledge model 124 in the processing graph of the storage device 120 . In addition, each process described below is a process that is executed at least before execution of the plan formulation by the plan formulation unit 113, and may be repeatedly executed for a certain period of time.

First, as a premise, the plan formulated by the planner is a plan that satisfies the planner, and the purpose of the plan formulation auxiliary device 100 is to output a plan similar to the plan formulated by the planner. Therefore, first, the planner prepares the input history 121 and the plan history 122 (step S501). For example, calculate The plan maker prepares a collection of input information as teacher data (for example, 1,000 pieces of input information in the past year) and adjusts the relaxation amount of the constraints as the plan maker's tacit knowledge based on the input information. A collection of planned planning information.

Next, the generation unit 111 of the plan preparation support device 100 inputs the input history 121 and the plan history 122 (step S502). More specifically, the generation unit 111 acquires the set of input information prepared in step S501 including restriction conditions, evaluation functions, and variables (for example, (Formula 1)) from the planner via the input device 140 as the input history 121, and store it in the storage device 120. In addition, the generation unit 111 acquires the set of plan information prepared by the planner prepared in step S501 as the plan history 122 and stores it in the storage device 120 .

Next, the generation unit 111 sorts the input history 121 acquired in step S502 by the same project number 201 (step S503). Here, each classified group is called "input information" in the following steps.

Next, the generation unit 111 selects the plan number 201 that has not been selected so far from the input history 121 acquired in step S502 (step S504).

Next, the generation unit 111 determines whether there is a plan number 201 that was not selected in step S504 (step S505). If the generation unit 111 determines that there is no unselected plan number 201 in step S504, the process proceeds to step S506. If it determines that there is no unselected plan number 201, the process proceeds to step S509. That is, the processes from step S506 to step S508 are repeatedly input only by the number of types of the plan number 201 of the history 121 .

When there is a plan number 201 that was not selected in step S505. Next, the generation unit 111 performs evaluation indexing using the relaxation amount of the restriction condition as a penalty (step S506). More specifically, the generation unit 111 extracts the restriction conditions, evaluation functions, and variables (for example, (Formula 1)) included in the input information equal to the project number 201 from the input history 121, and adds the relaxation of the restriction conditions The quantity is set as the evaluation index of the penalty rule, and in the subsequent steps, it is regarded as a multi-purpose optimization problem (for example, the following (Formula 2)). This is because there is more than one plan that relaxes the restriction conditions in the plan history 122 prepared by the plan maker.

Next, the generation unit 111 calculates one or more candidate plan information from the input information equal to the plan number 201 (step S507). More specifically, the generation unit 111 uses a general method (such as a local search method, a genetic algorithm, etc.) for solving a multi-purpose optimization problem (such as (Formula 2)) from the input information equal to the plan number 201. The calculation is called the Pareto optimal solution Multiple candidate plans (generate plan information). In multi-objective optimization problems, an executable solution that cannot improve multiple evaluation indicators at the same time is called a Pareto optimal solution. For example, a solution that maximizes value without violating constraints (relaxation amount is zero) The plan group that violates slightly and has higher value, and the plan group that has higher value but also violates a large amount is called Pareto optimal solution.

Next, the generation unit 111 stores the identifier 410 number, which is the identification information identifying the plan information in each piece of plan information (Pareto optimal solution) calculated in step S507, as the plan candidate 123 in the storage device 120, and processes Return to step S504 (step S508).

If there is no plan number 201 that was not selected in step S505, the generation unit 111 generates a set of feature vectors and labels for each piece of plan information from the plan history 122 (step S509). More specifically, the generation unit 111 reads the plan history 122 created by the planner from the storage device 120, and generates a feature vector describing the plan information for each plan number 301 (for example, the following (Equation 3) ), generate teacher information by assigning tags. It is expected that the elements that generally constitute a feature vector are elements that the planner considers intentionally or unintentionally when formulating the plan. An example of the feature vector is shown in (Formula 3).

For example, if the average value of the production cost of the input information is low and the unevenness (standard deviation) is small, and the average value of the product value is high and the unevenness (standard deviation) is small, the planner will set the amount of relaxation of the constraints on the production capacity. Minimization is prioritized over profit maximization. In addition, for example, when the number of products for which information is input is large, the unevenness (standard deviation) of production costs is large, and the unevenness (standard deviation) of product values is large, the planners develop a plan that prioritizes production capacity over maximizing profits. A plan that minimizes the amount of relaxation of constraints. Assume that such a planner formulates a plan. As shown in (Formula 3), the feature vector includes two types of features: a feature vector plus a feature of the input information and a feature vector plus a feature of the plan information. composed of vectors.

The number of products, the average and standard deviation of product production costs, and the average and standard deviation of product values are set to the feature vector that adds the characteristics of the input information, and the value of the evaluation index is set to the feature vector that adds the features of the plan information. The elements constituting the feature vector can be selected by the planner, etc., or each element can be prepared in advance and a collection of plan information of the plan history 122 created by the planner can be based on similarity using multiple regression analysis, cluster analysis, etc. Calculate effective elements and select these elements when classifying indicators. In addition, the generation unit 111 sets the value of the tag to “1” indicating a plan that the planner is satisfied with.

Next, the generation unit 111 generates for each plan information from the plan candidates 123 The set of feature vectors and labels (step S510). More specifically, the generation unit 111 reads the plan candidates 123 calculated in step S507 from the storage device 120 and generates a feature vector describing the plan information (for example, (Formula) for each plan number 301 and identifier 410 Sub-3)), assign tags to generate teacher information. The elements constituting the feature vector are the same as those set in step S509 described above.

Here, FIG. 6 is an image of two evaluation index axes plotted in (Equation 2) for each plan information, including the plan history 122 created by the plan maker and the plan candidates 123 calculated in step S507. Figure.

The left picture in Figure 6 represents the first plan information, and the right picture represents the i-th plan information. The horizontal axis is the reciprocal of profit (total product value), which is the first evaluation index of (Equation 2). The further to the left, the greater the profit. The vertical axis represents the relaxation amount (the value obtained by subtracting the maximum production capacity from the total production cost) as the second evaluation index of (Equation 2), and the relaxation amount decreases as the value goes downward. These two evaluation indicators are in a trade-off relationship, and there are multiple optimal plans (Pareto optimal solutions). This is Project Candidate 123, represented by a black dot in Figure 6 . With respect to the plan candidates 123, the plan most preferred by the selected planner (the plan formulated by the planner) is called the optimal solution, and is represented by a white circle in FIG. 6 .

In step S510 described above, the generation unit 111 assigns a label to each generated feature vector to generate teacher materials. At this time, the value of the label is set to "0" indicating that it does not satisfy the plan maker's plan.

Here, in order to solve the problem that the number of teacher materials with a tag value of “1” is overwhelmingly smaller than the number of teacher materials with a tag value of “0”, as shown in FIG. 7 or 8 , the calculation may be changed. Draw labels for candidate 123 projects value.

FIG. 7 is an image diagram in which the value of the label of the plan of the plan candidate 123 is changed from “1” to “0”. For example, the generation unit 111 acquires the optimal solution obtained in the above step S509, and the value of the label of the plan (nearby solution) of the plan candidate 123 that is close to the optimal solution is set to the same value as the optimal solution to indicate that it satisfies the planner. "1" for the plan, and the value of the label of the plan of the other plan candidates 123 is set to "0" indicating that it does not satisfy the plan of the plan creator. In addition, the close solution can be, for example, a plan of the plan candidate 123 that is within a certain range (for example, within a predetermined distance) from the optimal solution.

FIG. 8 is an image diagram in which the value of the label of the plan of the plan candidate 123 is changed from “1” to “(none)”. For example, the generation unit 111 acquires the optimal solution obtained in step S509 and excludes the teacher data (the set of feature vectors and labels) (including Delete.) (for example, data other than machine learning objects), and set the value of the plan label of the other plan candidates 123 to "0" indicating that it does not satisfy the plan maker's plan.

Next, the generation unit 111 machine-learns the set of feature vectors and labels generated in steps S509 and S510 as teacher materials, and generates a regression model as the knowledge model 124 (step S511). Here, machine learning uses usual methods (such as logistic regression, random forests, deep learning, etc.). (Formula 4) is an expression showing an example of the knowledge model 124.

[Formula 4] Optimal solution probability: P = f ( y ) f: knowledge model y: feature vector ...(Formula 4)

The knowledge model 124 is a function that outputs a probability as an optimal solution (hereinafter referred to as optimal solution probability) when a feature vector is input. Machine learning refers to generating this function. The optimal solution probability is a real number from "0" to "1". It means that when it is close to "0", the probability of being the optimal solution is low, and when it is close to "1", the probability of being the optimal solution is high. That is, it is expected that when a feature vector close to a feature vector generated from the plan information created by the planner is input to the knowledge model 124, a value close to “1” will be output. On the other hand, when a feature vector far from the feature vector generated from the plan information created by the planner is input to the knowledge model 124, a value close to “0” is output.

Here, when two different eigenvectors are close, it means that the Mahalanobis distance between the two eigenvectors is short, for example, and when two different eigenvectors are far apart, it means that the Mahalanobis distance between the two eigenvectors is long, for example.

Next, the output unit 112 stores the regression model generated in step S511 as the knowledge model 124 in the storage device 120 (step S512), and ends the process. In addition, after the knowledge model 124 is generated normally and stored in the storage device 120 , the planning assistance device 100 may notify the process through the output device 150 that the process has ended normally.

Here, using FIGS. 9 and 10 , before using the knowledge model 124 to describe the process of formulating a plan, during the series of processes from step S501 to step S512 , the planner browses through the output device 150 and uses the diagram. 11 explains the screen for inputting appropriate information through the input device 140.

FIG. 11 is a diagram showing an example of the knowledge model generation screen 1100.

The knowledge model generation screen 1100 includes a text input area 1101, an input history reading button 1102, a plan history reading button 1103, a knowledge model generation button 1104, and a teacher data distribution output area 1105.

The text input area 1101 is an area where the planner inputs the above-mentioned constraints, evaluation indicators, and variables. The input history read button 1102 is a button used by the planner to select a text file in which the contents of the input history 121 are recorded. The plan history read button 1103 is a button used by the plan maker to select a text file in which the contents of the plan history 122 are recorded. The knowledge model generation button 1104 is a button used by the planner to instruct the generation of the knowledge model 124 . The teacher data distribution output area 1105 is an area showing a pattern in which the teacher data (a set of feature vectors and labels) generated by the generation unit 111 is distributed with the evaluation index as an axis.

For example, the planner inputs (Formula 1) as the restriction condition, evaluation index, and variable into the text input area 1101 in the above step S502. In addition, for example, before starting step S502, the planner clicks the input history read button 1102 and selects a text file recording the contents of the input history 121 from the input storage device 120. In addition, for example, the planner clicks the plan history read button 1103 and selects a text file in which the contents of the plan history 122 are recorded from the storage device 120 .

Then, the planner clicks the knowledge model generation button 1104. The following steps are thus performed. As a result, the knowledge model 124 is accumulated in the storage device 120, and the teacher data as a set of all feature vectors and labels is displayed in the teacher data distribution output area 1105. In the case where there are multiple existing evaluation indicators and restricted conditions for admitting violations, the teacher data distribution output In Area 1105, two or three planners are selected from the existing evaluation indicators and constraints that are violated, and the distribution of teacher data is displayed with the selected evaluation indicators and constraints as the axis.

Next, as described above, the process of formulating a new plan based on new input information using the knowledge model 124 that adjusts the relaxation amount of the constraint as the tacit knowledge of the planner will be described.

FIG. 9 is a diagram illustrating an example of a processing sequence (mainly an example of a flowchart of processing executed by the plan-making assisting device 100) of the plan-making assisting method according to this embodiment. This flowchart is a diagram illustrating a process in which the plan formulation unit 113 generates new plan information similar to the plan information formulated by the planner based on the knowledge model 124 output by the output unit 112 .

In this flowchart, for example, it is assumed that the plan formulation assisting device 100 triggers the input device 140 to accept input information for the formulation of a new plan from the planner, activates the plan formulation unit 113, and starts processing. In addition, receiving new input information from the planner corresponds to information that has the same project as the record of the input history 121 illustrated in FIG. 2 and is summarized with the same project number.

First, the plan preparation unit 113 of the plan preparation support device 100 inputs input information (step S901). More specifically, the plan formulation unit 113 receives new input information including restriction conditions, evaluation functions, and variables (for example, (Formula 1)) from the planner via the input device 140 .

Next, the planning unit 113 indexes the relaxation amount evaluation of the restriction condition from the input information as a penalty (step S902). More specifically, the planning department 113 determines the restriction conditions, evaluation functions, etc. included in the new input information. Variables (for example, (Formula 1)), add the relaxation amount of the constraint as an evaluation index of a penalty, and treat it as a multi-purpose optimal problem (for example, (Formula 2)) in the following steps.

Next, the plan formulation unit 113 calculates one or more candidate plan information from the input information (step S903). More specifically, the planning unit 113 calculates a parameter called Pa from the new input information using a common method (such as a local search method, a genetic algorithm, etc.) for solving a multi-purpose optimal problem (such as (Formula 2)). Multiple project information for candidates of the reto-optimal solution.

Next, the plan formulation unit 113 generates a feature vector for each piece of candidate plan information (step S904). More specifically, the plan formulation unit 113 generates the plurality of candidate plan information calculated in step S903 and a feature vector describing the plan information for each piece of plan information (eg (Formula 3)). The elements constituting the feature vector are the same as the elements set in step S509 described above.

Next, the plan formulation unit 113 uses the knowledge model 124 to calculate the optimal solution probability for each candidate plan information (step S905). More specifically, the plan formulation unit 113 reads the knowledge model 124 from the storage device 120 and inputs the feature vector generated for each piece of plan information that became a candidate in step S904 to the knowledge model 124 (for example (Equation 4) )) and calculate the best solution probability.

FIG. 10 is an image diagram showing the optimal solution probability for each of the five candidate plan information pieces. In Figure 10, among the five plan information, the one with the highest profit (the value of f ₁ is the smallest) P ¹ is "0.1", and the one with the smallest relaxation (the value of f ₂ is the smallest) P ⁵ is "0.3", and the probability of the best solution is low. , the profit is high and the amount of relaxation is small, P ³ is "0.9", and the probability of the best solution becomes higher.

Next, the plan formulation unit 113 outputs the plan data with the highest optimal solution probability. message (step S906). More specifically, the plan making unit 113 selects the plan information with the largest value among the optimal solution probabilities calculated in step S905, and outputs the selected plan information to the plan maker via the output device 150, and ends. Processing by the planning department 113.

Here, during the series of processes from step S901 to step S906 described above, the planner views the screen through the output device 150 and inputs appropriate information through the input device 140 .

FIG. 12 is a diagram showing an example of the plan creation screen 1200.

The plan preparation screen 1200 includes: an input information read button 1201, an input information text output field 1202, a knowledge model read button 1203, a plan information generation button 1204, a plan information output area 1205, and a teacher data distribution output area 1206.

The input information read button 1201 is a button for the plan maker to select a text file in which the contents of the input information necessary for making the plan are recorded. The input information text output field 1202 is a field that displays the read input information. The knowledge model read button 1203 is a button for the planner to select the knowledge model 124 generated by the generation unit 111 . The plan information generation button 1204 is a button used by the planner to activate the plan formulation unit 113 and instruct the generation of plan information (formulation of a plan). The plan information output area 1205 is an area where the plan information output by the plan formulation unit 113 is displayed in a table format. The teacher information distribution output area 1206 is an area showing a pattern in which teacher information (a combination of feature vectors and labels) similar to the plan information output by the plan formulation unit 113 is distributed around the evaluation index.

For example, in the above step S901, the planner inputs by clicking The information read button 1201 is used to select a text file recording the content of the input information from the storage device 120 to obtain the input information used to formulate a new plan. In addition, by clicking the knowledge model read button 1203, the knowledge model 124 used for the formulation of a new plan is retrieved from the storage device 120, and the plan information generation button 1204 is clicked.

As a result, the following steps are executed, and the plan information newly generated as a result is displayed in the plan information output area 1205 . In addition, teacher information as a group of feature vectors and labels similar to the generated plan information is displayed on the teacher information distribution output area 1206 . When there are multiple existing evaluation indicators and restricted conditions for admitting violations, you can also select 2 or 3 planners from the existing evaluation indicators and restricted conditions for admitting violations in the teacher information distribution output area 1206. , display the distribution of teacher data with the selected evaluation indicators and constraints as the axis.

As mentioned above, in this embodiment, a knowledge model that adjusts the relaxation amount of the restriction condition is generated from the input history and the plan history. As the tacit knowledge of the plan maker, it is reflected in the formulation of future plans, and the satisfaction level can be output. High plans. This makes it possible to formulate an effective plan that reflects tacit knowledge that adjusts the evaluation index and the relaxation amount of the restriction conditions based on the input information.

(2) Second embodiment

This embodiment will be described. Among these, differences from the first embodiment will be mainly described. In this embodiment, the planning assistance device 100 shown in FIG. 1 effectively formulates feedback when there are multiple types of tacit knowledge that adjusts evaluation indicators and relaxation amounts of constraints based on input information. Different project forms that reflect this tacit knowledge. The difference in tacit knowledge may be, for example, the difference in tacit knowledge of each plan maker when there are multiple plan makers, or the difference in tacit knowledge of one plan maker that changes with time. different.

The generation unit 111 of the planning support device 100 of the present embodiment executes the same process as the process of the generation unit 111 of the first embodiment ( FIG. 5 ) except for a part. Therefore, in the following, the processing of the generation unit 111 in this embodiment that is different from the first embodiment will be described.

In the process illustrated in FIG. 5 , when preparing the set of input information and the set of plan information in step S501 , the generation unit 111 of this embodiment divides and prepares the sets for each planner. Alternatively, you can prepare it by dividing it into the last year, the last month, and the last week.

Next, in the process illustrated in FIG. 5 , the generation unit 111 of this embodiment inputs the set of input information and the set of plan information divided in the above-described step S501 via the input device 140 in step S502 to obtain the divided set of input information. The processing of the generation unit 111 is executed for the number of .

Through the above processing, the generation unit 111 of this embodiment generates the knowledge model 124 in the number divided in step S501.

In addition, the planning unit 113 of this embodiment executes the same process as the process of the plan unit 113 of the first embodiment ( FIG. 9 ) except for a part. Therefore, in the following, the processing by the planning unit 113 of this embodiment that is different from that of the first embodiment will be described.

In the process illustrated in FIG. 9 , the planning unit 113 of the present embodiment extracts from the storage the quantity divided in the above-mentioned step S501 in step S905 . The storage device 120 reads the knowledge model 124, and inputs the feature vector generated for the planning information that became the candidate in step S904 into the knowledge model 124 (for example, (Formula 4)) to calculate the optimal solution probability based on the number of divisions.

Next, in the process illustrated in FIG. 9 , in step S906 , the planning unit 113 of the present embodiment selects the optimal solution probabilities calculated in step S905 by the number divided in step S501 . The plan information with the largest value is output to the plan maker via the output device 150, and the processing of the plan making unit 113 is completed.

As described above, according to this embodiment, for example, when there are multiple planners, the tacit knowledge of each planner is different, or the tacit knowledge of one planner changes with the passage of time. When there are multiple types of tacit knowledge such as the difference in knowledge, multiple knowledge models that adjust the relaxation amount of the restriction conditions are generated from the input history and the plan history, and these tacit knowledge are used as the tacit knowledge of the planner. The difference in sexual knowledge is reflected in the formulation of future plans, and plans with a high degree of satisfaction can be output to each knowledge model. This enables effective planning reflecting tacit knowledge that adjusts the evaluation index and the relaxation amount of the restriction conditions based on the input information.

As mentioned above, the preferable mode etc. for carrying out this invention were demonstrated concretely, However, this invention is not limited to this, Various changes are possible within the range which does not deviate from the summary.

(3) Other embodiments

In addition, in the above-mentioned embodiment, the case where the present invention is applied to a planning assistance device has been described, but the present invention is not limited to this and may It can be widely used in various other systems, devices, methods, and programs.

In addition, in the above embodiment, the structure of each table is an example, and one table may be divided into two or more tables, and all or part of two or more tables may be one table.

In addition, in the above-mentioned embodiment, for convenience of explanation, various data are described using XX tables and XX folders, but the data structure is not limited and may also be expressed as XX information, etc.

In addition, in the above description, information such as programs, tables, and folders that implement each function can be stored in storage devices such as memory, hard disks, and SSDs, and recording media such as IC cards, SD cards, and DVDs.

The above-described embodiment has, for example, the following characteristic structures.

The plan preparation assistance device (for example, the plan preparation assistance device 100) includes a generation unit (for example, the generation unit 111) based on an input history (for example, the input history 121) including constraints and evaluation indicators in each plan and information on Information indicating the relaxation amount of the amount that violates the above-mentioned restriction conditions of each of the above plans (for example, the relaxation amount 309, the maximum production capacity 303 as the information used to calculate the relaxation amount 309, the production cost 305, the selection flag 307, etc.) drawing history (for example, plan history 122), generating a knowledge model (for example, knowledge model 124, (Formula 4)) representing the relationship between the relaxation amount of the restriction condition and the above evaluation index; and an output unit (for example, output unit 112), It outputs the above-mentioned knowledge model so that the plan-making part (for example, the plan-making part 113) who wants to use the above-mentioned knowledge model to make a plan can use it (for example, store it in the storage device 120, notify it to the plan-making part 113, and send it to other computer, display on the output device 150, etc.) the above knowledge model.

In the above structure, a knowledge model representing the relationship between the evaluation index and the relaxation amount of the restriction condition is generated. Based on the knowledge model, for example, the planning department can grasp the relationship (equilibrium) of how much violation the planner allowed based on how much value there was in the past. Therefore, for example, the planning department can formulate a plan that reflects the equilibrium close to the past plan based on the input information, that is, a plan that reflects the relationship between the evaluation index and the relaxation amount of the restriction condition.

The above-mentioned generation unit generates, for each of the above-mentioned plans, based on the above-mentioned plan history, the feature vector of each of the above-mentioned plans (for example, (Formula 3)) and the first label indicating that it was created by the plan maker (for example, "1 ”) group, based on the above-mentioned input history, the above-mentioned each plan is optimized with the above-mentioned evaluation index and the evaluation index that minimizes the above-mentioned relaxation amount to generate a plan candidate (for example, plan candidate 123), and generate the above-mentioned plan candidate A set of feature vectors of each plan and a second label (for example, “0”) indicating that the plan is a candidate, and machine learning is performed on the generated set of feature vectors and labels to generate the above-mentioned knowledge model.

According to the above structure, by comparing the teacher data of the correct solution (the set of feature vectors of the project history and the first label) and the teacher data of the incorrect solution (the set of feature vectors of the project candidates and the second tag) Learning at the same time can improve the accuracy of the knowledge model. This allows the planning department to formulate plans that are more satisfying to the planners.

In addition, when the knowledge model is generated using the plans of the plan history instead of the plans of the plan candidates, the generation of the knowledge model can be simplified and the generation time of the knowledge model can be shortened. In addition, the generation method may also be determined based on the skewness of the distribution of the project feature vector in the project history and the first label group. information when forming a knowledge model. For example, when the standard deviation of the distribution exceeds the threshold, it is decided to generate a knowledge model for the usage plan history and plan candidates, and when the standard deviation of the distribution does not exceed the threshold, it is decided to generate a knowledge model for the usage plan history.

The generation unit generates the knowledge model by changing the label of the second label group located near the first label group to the first label (see, for example, FIG. 7 ).

According to the above structure, by combining the teacher data of the correct answer (the set of feature vectors of the project history and the first label) to the teacher data of the incorrect answer (the set of feature vectors of the project candidates and the second tag) ) to correct teacher materials can avoid a situation where there are overwhelmingly few correct teacher materials. This allows, for example, the generation unit to secure correct answer teacher materials, thereby improving the accuracy of the knowledge model.

The generation unit generates the knowledge model by excluding the second label group located near the first label group for each of the plans (see, for example, FIG. 8 ).

According to the above structure, by excluding the teacher data of the correct answer (the set of feature vectors of the project history and the first label) that are nearby the non-correct answer teacher data (the set of feature vectors of the project candidate and the second label) ), it is possible to avoid a situation in which there are overwhelmingly few teacher materials with correct answers. This allows, for example, the generation unit to balance the number of correct teacher materials and incorrect teacher materials, thereby improving the accuracy of the knowledge model.

The above-mentioned generating unit generates data based on pre-specified conditions (such as temperature, humidity, weather, air pressure, wind intensity and other phenomena, seasons, time zones and other periods, days). (period time and other time, days of the week, holidays, working days, etc.), for each of the above-mentioned plans, the feature vector of each of the above-mentioned plans is generated based on the above-mentioned plan history, and the first feature vector representing the situation set by the plan maker is generated. A group of labels, and machine learning is performed on the generated feature vector and the group of labels to generate the above knowledge model.

In the above structure, the relationship between the evaluation index and the relaxation amount of the restriction condition is generated based on the knowledge model expressed by the condition. Therefore, the plan formulation unit divides the input information according to the predetermined conditions, thereby making it possible to formulate a plan corresponding to the input information. . For example, the planning department can determine the season (summer, winter, etc.) from the date and time data included in the input information, and can formulate a plan corresponding to the season.

The above-mentioned generation unit generates data for each pre-specified point of view (the pre-specified point of view is a point of view required for the plan maker's plan. For example, it indicates the type of tacit knowledge, the plan maker, the period of one year, one month, one week etc.), for each of the above-mentioned projects, a feature vector of each of the above-mentioned projects and a group representing the first label created by the plan maker are generated for each of the above-mentioned projects based on the above-mentioned project history, and the generated set of feature vectors and labels are generated by machine learning. The above knowledge model.

In the above structure, a knowledge model is generated for each pre-specified viewpoint. For example, when a knowledge model is generated for a period of one year, one month, or one week, the planner can obtain a knowledge model that reflects the period of one year, one month, or one week by selecting the knowledge model for any one period. Tendency plan. In addition, for example, when a knowledge model is generated for each planner, the planner can select his or her own knowledge model to obtain a plan that reflects his or her own tacit knowledge, or by selecting the knowledge of technicians. Models can be used to obtain plans that reflect the tacit knowledge of technicians.

In addition, in the above structure, appropriate changes, rearrangements, combinations, and omissions may be made within the scope that does not deviate from the gist of the present invention.

100: Planning assistance device

111:Generation Department

112:Output Department

113: Planning Department

120:Storage device

121:Input history

122:Project History

123:Project candidate

124:Knowledge model

125:Program

130:Memory

140:Input device

150:Output device

Claims (4)

A plan preparation auxiliary device, including: a generation unit that generates data from an input history including constraints and evaluation indicators in each plan, and includes an amount related to a relaxation amount indicating a violation of the aforementioned constraints of each of the plans. The plan history of the information is based on the aforementioned plan history. For each of the aforementioned plans, a feature vector of each of the aforementioned plans is generated, and a group of first tags indicating the situation formulated by the plan maker is based on the aforementioned input. Historically, for each of the aforementioned plans, optimization of the aforementioned evaluation index and the evaluation index that minimizes the aforementioned relaxation amount is performed to generate plan candidates, and the feature vectors of each plan and the sum of the feature vectors of the aforementioned plan candidates are generated to indicate that the plan candidates are The group of second tags of the situation, through this, generates a knowledge model representing the relationship between the relaxation amount of the aforementioned restriction conditions and the aforementioned evaluation index; and an output unit that outputs the aforementioned knowledge model so that the planning department can use it, The planning department uses the aforementioned knowledge model to formulate plans. In the planning assistance device of claim 1, the generating unit generates the knowledge model by changing the tags of the second tag group located near the first tag group to the first tag. In the plan preparation assisting device of claim 1, the generating unit generates the knowledge model by excluding the second tag group located near the first tag group for each of the plans. An auxiliary method for plan formulation, including: a first step, in which the generation part obtains input history including constraints and evaluation indicators in each plan, and includes and indicates violations of each of the aforementioned plans. The plan history of the information related to the amount of relaxation of the aforementioned restriction conditions, based on the aforementioned plan history, generates the feature vector of each of the aforementioned plans for each of the aforementioned plans, and represents the situation formulated by the plan maker. The first label group of , based on the aforementioned input history, optimizes the aforementioned evaluation index and the evaluation index that minimizes the aforementioned relaxation amount for each of the aforementioned plans to generate plan candidates, and generate each plan of the aforementioned plan candidate The feature vector of , and a group of second tags representing events that are project candidates are used to generate a knowledge model representing the relationship between the relaxation amount of the aforementioned restriction conditions and the aforementioned evaluation index; and a second step, which is the output part The aforementioned knowledge model is output to be used by the planning department, and the planning department uses the aforementioned knowledge model to formulate plans.

Images