JP2021189933A - Planning system, and planning apparatus and planning method therefor - Google Patents

Abstract

To develop a plan with high evaluation by considering a relationship between the relaxation amount of constraint and a plan evaluation value for a plan requiring relaxation of constraint conditions.SOLUTION: A planning apparatus includes a processor and a storage device and develops a production plan. The processor includes a planning section for developing a plan on the basis of constraint conditions and an evaluation function created from input order information and facility information to be planned and a constraint relaxation section for relaxing violation constraints when constraint conditions are violated. The constraint relaxation section identifies constraint relating to the violation constraints, creates a relaxing pattern, extracts a relationship between a relaxation amount and an evaluation value with an evaluation function and an application frequency of the relaxation amount from historic data for each the relaxation pattern, selects a relaxation amount on the basis of the relationship between the relaxation amount and the evaluation value and the application frequency of the relaxation amount. The planning section develops a plan on the basis of the relaxation amount for each developed relaxation pattern, evaluates the plan with the relaxation function and adopts a plan with high evaluation.SELECTED DRAWING: Figure 1

Description

The present invention relates to a planning system in consideration of relaxation of constraints.

In the planning system, when planning is performed, there are many constraints in the actual problem, and it is difficult to satisfy all the constraints. Therefore, in the planning system, planning by efficient relaxation of constraints when all the constraints are not satisfied has become a general issue.

Patent Document 1 is a background technique in this technical field. Patent Document 1 is a delivery planning system for drafting a delivery plan for goods, and holds constraint conditions for drafting a delivery plan and cost information used for calculating a delivery cost based on the delivery plan. A delivery plan and a cost calculation unit that formulate a delivery plan before relaxation of the constraint condition so as to satisfy the constraint condition, and calculate the delivery cost of the delivery plan before relaxation based on the cost information. And a constraint relaxation unit that generates a constraint relaxation plan by relaxing the constraint condition, and the delivery planning and cost calculation unit further relaxes the constraint condition so as to satisfy the constraint relaxation plan. When the delivery cost of the later delivery plan is drafted, the delivery cost of the post-mitigation delivery plan is calculated based on the cost information, and the delivery cost of the post-mitigation delivery plan is lower than the delivery cost of the pre-mitigation delivery plan. The point of outputting the restriction relaxation plan is described.

International Publication No. 2017/187482

In Patent Document 1, one or more constraint relaxation plans are generated by relaxing to one or more relaxation amounts based on the step size of relaxation, and a constraint relaxation plan lower than the delivery cost is output. However, it was difficult to construct an efficient mitigation plan because the relationship between the mitigation amount and the delivery cost was not quantitatively considered and it was not possible to know how many mitigation plans should be prepared in advance.

The present invention is, for example, a planning device having a processing device and a storage device to formulate a production plan, and the processing device is a constraint generated from input order information and equipment information of the planned schedule. It has a planning department that creates a plan based on conditions and evaluation functions, and a constraint relaxation unit that relaxes the violation constraint when the constraint condition is violated. A relaxation pattern is generated, and for each relaxation pattern, the relationship between the relaxation amount and the evaluation value in the evaluation function and the application frequency of the relaxation amount are extracted, and the relationship between the relaxation amount and the evaluation value and the application of the relaxation amount. The mitigation amount is selected based on the frequency, and the planning department formulates a plan based on the mitigation amount for each mitigation pattern generated, evaluates it with an evaluation function, and adopts a better plan for evaluation.

According to the present invention, it is possible to formulate a plan with good evaluation in consideration of the relationship between the amount of relaxation of the constraint and the evaluation value of the plan.

It is a block diagram of the planning system in an Example. It is a functional block diagram of the planning part in an Example. It is a functional block diagram of the constraint relation extraction part in an Example. It is a functional block diagram of the relaxation pattern classification part in an Example. It is a functional block diagram of the relaxation information extraction part in an Example. It is a functional block diagram of the relaxation amount application part in an Example. It is a hardware block diagram of the planning apparatus in an Example. It is a processing flow diagram of the planning apparatus in an Example. It is a data structure diagram of the order information DB in an Example. It is a data structure diagram of the equipment information DB in an Example. It is a processing flow diagram of constraint relaxation in an Example. It is a figure which shows the constraint priority in an Example. It is a figure which shows the constraint-related information in an Example. It is a figure which shows the relaxation amount and evaluation value distribution in an Example. It is a figure which shows the relaxation amount application frequency in an Example. It is a figure which shows the display screen of the planning system to be displayed on the user terminal in an Example.

Hereinafter, examples of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a configuration diagram of a planning system in this embodiment. In FIG. 1, the planning system is composed of a planning device 100, an order information DB (database) 101, equipment information DB 102, and a user terminal 105.

The order information DB 101 is a device for storing an order to be produced. The order information includes information such as the type, quantity, and delivery date of the product requested by the customer.

The equipment information DB 102 is a device that stores descriptions such as the production capacity of the equipment, the specifications of the products to be produced, and the differences in the specifications of the products lined up before and after in one manufacturing line.

The planning device 100 has a planning unit 103, a constraint relaxation unit 104, and a planning result DB 106 as its functional units.

The planning unit 103 reads the order information of the planned schedule from the order information DB 101 and the information of the equipment that can be used from the equipment information DB 102, and creates a plan according to the decision variables, the constraints, and the evaluation function that evaluates the quality of the plan. Is.

The constraint relaxation unit 104 is a function of relaxing the violation constraint and sending the relaxation amount of the violation constraint to the planning unit 103 when all the constraint conditions cannot be satisfied in the planning by the planning unit 103. The constraint relaxation unit 104 is further composed of a constraint relevance extraction unit 111, a relaxation pattern classification unit 112, a relaxation information extraction unit 113, and a relaxation amount application unit 114 as its functional units.

The plan result DB 106 is a device for storing the automatic plan created by the plan planning unit 103. The user terminal 105 is a terminal operated by the user. Although the planning device 100 will be described as having the planning result DB 106, the planning result DB 106 is arranged outside the planning device 100, and the planning device 100 accesses the external planning result DB 106 by the network interface described later. You may do it.

FIG. 2 is a functional configuration diagram of the planning unit 103 in this embodiment. In FIG. 2, 201 is a decision variable generation unit, 202 is a constraint condition generation unit, 203 is an evaluation index generation unit, 204 is a planning planning unit, and 205 is a planning result output unit.

The decision variable generation unit 201 is a processing unit that sets a range of production in the equipment. The constraint condition generation unit 202 is a processing unit that reads the order information of the planned schedule from the order information DB 101 and the equipment information from the equipment information DB 102, and creates constraint conditions such as the maximum production amount of the equipment and the delivery date of the ordered product. The evaluation index generation unit 203 is a processing unit that evaluates a plan based on the production cost of the product produced in one plan and creates an evaluation function.

The planning unit 204 is a processing unit that determines the numerical value of the decision variable by the optimization algorithm based on the decision variable generation unit 201, the constraint condition generation unit 202, the decision variable generated from the evaluation index generation unit 203, the constraint condition and the evaluation function. Is.

The plan result output unit 205 is a processing unit that determines the end conditions of the plan such as the plan accuracy and the processing time, causes the plan planning unit 204 to be replanned, or ends the planning process and outputs the plan result and the like. Further, when the constraint condition is violated, the plan result output unit 205 includes information on the decision variable, the constraint condition, and the evaluation function generated from the decision variable generation unit 201, the constraint condition generation unit 202, and the evaluation index generation unit 203. This is a processing unit that sends the violation constraint information output from the planning unit 204 to the constraint relaxation unit 104.

FIG. 3 is a functional configuration diagram of the constraint relevance extraction unit 111 in the constraint relaxation unit 104. In FIG. 3, the related constraint condition extraction unit 301 is a processing unit that specifies the constraint condition related to the violation constraint based on the constraint condition, the evaluation function information, and the violation constraint information output from the planning unit 103. ..

The constraint condition priority extraction unit 302 is a processing unit that extracts the degree of violation of the related constraint condition from the planning result DB 106 for the related constraint condition specified from the related constraint condition extraction unit 301 and sets it as a priority.

The constraint-related information creation unit 303 combines and outputs the related constraint information and the priority information based on the related constraint condition and the constraint condition priority extracted from the related constraint condition extraction unit 301 and the constraint condition priority extraction unit 302. It is a processing unit.

FIG. 4 is a functional configuration diagram of the relaxation pattern classification unit 112 in the constraint relaxation unit 104. In FIG. 4, the relational constraint combination creation unit 401 is a processing unit that creates a relational constraint combination based on the relational constraint information and priority information extracted from the relational constraint condition extraction unit 301.

The relaxation pattern creation unit 402 calculates the frequency of the relation constraint combination from the planning result DB 106 for the relation constraint combination created from the relation constraint combination creation unit 401, and adopts and generates the relation constraint combination with high frequency as the relaxation pattern. It is a processing unit that outputs.

FIG. 5 is a functional configuration diagram of the relaxation information extraction unit 113 in the constraint relaxation unit 104. In FIG. 5, the similarity plan extraction unit 501 is a processing unit that extracts similar constraints from the history data of the plan result DB 106 for each relaxation pattern generated from the relaxation pattern classification unit 112.

The evaluation function distribution creation unit 502 is a processing unit that extracts and creates a distribution of the relaxation amount of the constraint condition in the relaxation pattern and the evaluation value in the evaluation function of the corresponding plan for the similarity plan extracted from the similarity plan extraction unit 501. Is.

The relaxation amount application frequency creation unit 503 is a processing unit that creates a relationship between the relaxation amount of the constraint condition in the relaxation pattern and the application frequency of the corresponding relaxation amount for the similar plan extracted from the similarity plan extraction unit 501.

The relaxation information output unit 504 is a processing unit that outputs the distribution of the relaxation amount and the evaluation value and the application frequency of the relaxation amount created by the evaluation function distribution creation unit 502 and the relaxation amount application frequency creation unit 503.

FIG. 6 is a functional configuration diagram of the relaxation amount application unit 114 in the constraint relaxation unit 104. In FIG. 6, the mitigation amount candidate selection unit 601 selects a mitigation amount candidate based on the mitigation amount created by the mitigation information extraction unit 113 and the change in the distribution of the evaluation values in the evaluation function of the corresponding plan. It is a department.

The application frequency candidate selection unit 602 is a processing unit that selects a mitigation amount candidate with a high frequency of use based on the relationship between the mitigation amount created from the mitigation information extraction unit 113 and the application frequency of the corresponding mitigation amount.

The relaxation amount candidate output unit 603 is a processing unit that adopts one relaxation amount candidate based on the candidates selected from the relaxation amount candidate selection unit 601 and the application frequency candidate selection unit 602 and outputs the relaxation amount candidate to the planning unit 103. ..

FIG. 7 is a hardware configuration diagram of the planning device 100 in this embodiment. In FIG. 7, the planning device 100 is composed of a general PC system. That is, it is composed of a CPU (processing device) 701, a memory (storage device) 702, an interface 703, a network interface 704, a keyboard 705, a screen 706, a mouse 707, and a hard disk 708.

The CPU 701 is a central processing unit, and executes the above-mentioned various functions by executing a program recorded in the memory 702 or a program previously transferred from the hard disk 708 to the memory 702. If necessary, the program can be used by a PC and may be introduced by a removable storage medium. In this case, a device for reading the storage medium is connected to the interface 703. As such a storage medium and a device for reading the storage medium, those using an optical disk and those using a flash memory are generally known. Further, the program may be introduced into the PC by the network interface 704 via the communication medium (communication line or carrier wave on the communication line), if necessary.

The memory 702 temporarily records programs and data. The interface 703 is for connecting the devices in the PC system, and each device in the PC system is connected.

The network interface 704 is a device for a PC outside the PC system to communicate with another database or the like, and communicates with the order information DB 101, the equipment information DB 102, and the user terminal 105.

The keyboard 705 is a device operated by the operator of the PC system in order to input commands and data to the PC system. The screen 706 is a device for displaying a processing result or the like. The mouse 707 is a device that specifies a position on the screen by moving a pointer displayed on the screen and causing an operator to press a button at an arbitrary place, and conveys some action to the CPU 701. It can also be replaced by a touch panel. The hard disk 708 is a device for storing programs and data.

Although the planning device 100 has been described as being composed of a general PC system, it is not necessary that all the functions in the planning device 100 are configured in one PC system, and a part of the functions thereof. May be configured by another PC system, or the planning device 100 may be configured by a plurality of PC systems.

FIG. 8 is a flowchart showing an example of the planning process in this embodiment.

In FIG. 8, step S801 is a process in which the user terminal 105 displays the order information from the order information DB 101 and the equipment information from the equipment information DB 102. In addition, it is a process of making the user confirm the order information. Here, it is assumed that the communication between the DB and the user terminal 105 can be executed by general communication, RPC (Remote Procedure Call), or the like, and hereinafter, the device-to-device communication is performed by such a method.

FIG. 9 is a diagram showing the data structure of the order information DB 101 in this embodiment. The order information DB 101 includes an order number 901, a product type 902, a delivery date 903, and a required quantity 904. The value of order number 701 is an identification number that uniquely identifies the order. The product type 902 is identification information that uniquely identifies the type of the product specified in the corresponding order. The value of the delivery date 903 represents the deadline for delivering the product for the corresponding order. For example, the value of the delivery date 903 represents the number of days until shipment. The required amount 904 is a value representing the number of products. The above items may include other data related to the order as the data necessary for creating the plan.

FIG. 10 is a diagram showing the data structure of the equipment information DB 102 in this embodiment. The equipment information DB 1001 includes equipment number 1001, manufactured product type 1002, maximum production amount 1003, and maintenance date 1004. The equipment number 1001 is an identification number that uniquely identifies the equipment. The manufactured product type 1002 is identification information that uniquely identifies the type of product that can be manufactured by the relevant equipment. The value of the maximum production amount 1003 represents the maximum production amount for the corresponding equipment and the corresponding manufactured product type. The maintenance date 1004 represents the scheduled maintenance date of the corresponding equipment. The above items may include other data on the equipment as data necessary for planning.

The above-mentioned order information and equipment information may include other data regarding workers and resources as data necessary for planning.

In step S802 of FIG. 8, the planning unit 103 sets the decision variable, the constraint condition and the evaluation function based on the order information, the equipment information, and the relaxation amount from the relaxation amount application unit 114, and the constraint condition is determined by the optimization algorithm. Optimize the evaluation function and make a plan while satisfying.

Here, there is the following setting method as an example of generation of a decision variable, a constraint condition, and an evaluation function.

Indexing of decision variables:
t: An index indicating the production date. As the range of t, for example, when indicating a production date of one month, 1 to 30 are set in the program of the planning unit 103.
i: Index indicating the variety. As for the range of i, for example, when the number of all varieties in the factory is 100, 1 to 100 are set in the program of the planning unit 103.
m: Index indicating equipment. As for the range of m, for example, when all the equipments in the factory are 10 units, 1 to 10 are set in the program of the planning unit 103.

Setting the coefficient of determination:
x _tim : A variable indicating the production amount of the variety i in the equipment m on the production date t. x _tim is set in the program of the planning unit 103.

Parameter settings:
C _im : A parameter indicating the maximum production amount of the variety i in the equipment m. The value of C _im is planning unit 103 reads the value of maximum production 1003 facilities m and varieties i from the facility information DB 102.
D _ti : A parameter indicating the required amount of the variety i on the production date t. As for _{the value of D ti,} the planning unit 103 reads the value of the production date t and the required quantity 904 of the product type i from the order information DB 101.
w _tim: parameter indicative of the cost of production at the facility m of varieties i in the production date t. to set the w _tim to the program of the planning unit 103. Alternatively, it may be read from external data.

Constraint settings:
The constraint that the production amount of the variety A _{must exceed the required amount D 30A} by C1: 30 is set by the formula (1).

C2: The constraint that the production amount of the variety A _{must exceed the required amount D10A} by 10 days is set by the formula (2).

The constraint that the production amount of the variety B _{must exceed the required amount D30B} by C3: 30 is set by the formula (3).

C4: The constraint that the production amount of the variety A _{must exceed the required amount D10B} by 10 days is set by the formula (4).

C1, C2, C3, and C4 are set in the program of the planning unit 103.

Evaluation function settings:
F: The function for calculating the production cost is set by the equation (5).

F is set in the program of the planning unit 103.

There are generally available mathematical planning software for determining variables, constraints and evaluation functions, and planning algorithms, which can also be used. In this embodiment, constraint programming is used. Regarding these calculations, the input in this embodiment is the constraint condition of the delivery date, quantity, product type, and target equipment of the order. Further, the output in this embodiment is the start / end time in the target equipment that observes the constraint conditions of the target equipment. As for the method of determining the start / end time, for example, a search method called backtracking is used. Backtracking is one of the general solution (planning candidate) search algorithms, and when finding a certain solution, the possible procedures are tried in order. In that procedure, if a solution is found, it is saved as a solution, and when it is found that a solution cannot be found, it returns to the previous state and tries another procedure. In the backtracking search method, a plurality of solutions (plan candidates) can be searched. Multiple plan candidates are evaluated by the evaluation function, and the plan candidates with high evaluation are adopted and used as the plan result. Further, when all the constraints are not satisfied, the violation constraint can be specified by sequentially adding the constraints C1 to C4. The evaluation function can be defined by, for example, a mathematical formula as shown in the formula (5).

The plan may be searched by defining decision variables, constraints and evaluation functions by another method such as linear programming. In that case, the planning unit 103 executes the search algorithm and formulates a plan.

In step S803 of FIG. 8, the planning unit 103 determines that the constraint condition is satisfied. If all the constraints can be satisfied, the process proceeds to step S804. When the constraint condition is violated, the planning unit 103 sends the defined decision variable, the information of the constraint condition and the evaluation function, and the information of the violated constraint condition to the constraint relaxation unit 104, and proceeds to step S806.

In step S804, the planning unit 103 displays the planning result and the restriction relaxation information in the case of relaxation on the user terminal 105, and causes the user to select and confirm.

In step S805, the user terminal 105 saves the planning result selected by the user in the planning result DB 106.

In step S806, the constraint relaxation unit 104 uses the planning result DB 106 based on the information of the decision variable, the constraint condition and the evaluation function, and the information of the violated constraint condition to determine the relaxation amount of the violated constraint, and the planning unit. Send to 103.

FIG. 11 is a flowchart showing an example of the constraint relaxation process in this embodiment. In FIG. 11, in step S1101, the constraint relevance extraction unit 111 in the constraint relaxation unit 104 determines that the violation constraint is based on the decision variable, the constraint condition and evaluation function information, and the violation constraint information output from the planning unit 103. The related constraint is specified, the priority of the related constraint is extracted from the planning result DB 106, and the constraint-related information is created according to the priority.

Here, an example of a method for identifying a constraint related to a violation constraint will be described. For example, assume that the constraint condition C1 is violated. All constraints are checked for the C1 decision variable x _tAm and the parameter D _tA , and constraints with common index decision variables and parameters are picked up. For example, in C2, x _tAm up to 10 days is common with C1. In C3, the equipment m is shared with C1 until the 30th. In C4, the equipment m is shared with C1, C2, and C3 until the 10th. Therefore, C2, C3, and C4 are related constraints of C1.

FIG. 12 shows an example of extracting the priority of the related constraint from the plan result DB 106 in this embodiment. In FIG. 12, the constraint condition identification ID 1201 is an ID that identifies the constraint condition of the planning unit 103. The degree of violation 1202 is the degree of violation using the calculation function of the constraint condition generation unit 202 of the planning unit 103 with respect to the plan of the plan result DB 106. The priority 1203 is the order in which the violation degree 1202 is in ascending order. The priority of the constraint may be set by the user.

FIG. 13 shows an example of constraint-related information in this embodiment. The constraint-related information is information that combines the related constraint information and the priority based on the related constraint and the constraint priority. Constraint-related information can be shown in the graph structure and tabular format shown in FIG.

Next, in step S1102 of FIG. 11, the mitigation pattern classification unit 112 creates a violation combination of related constraints, calculates the combination ratio from the planning result DB 106, and creates a combination with a high ratio as a mitigation pattern.

Here, examples of the violation combination include a combination in which there is no violation of C1 and only C4 is violated, or there is no violation of C1 and only C4 and C2 are violated. The target planning result is searched from the planning result DB 106 under the condition of the target combination, and the ratio of each combination is calculated. For example, a combination of two high proportions is defined as a relaxation pattern of C4 relaxation and C4 and C2 relaxation.

Next, in step S1103 of FIG. 11, the mitigation information extraction unit 113 creates the mitigation amount, the distribution of the evaluation value, and the application frequency of the mitigation amount from the target plan for each mitigation pattern, and outputs the mitigation information as mitigation information.

FIG. 14 shows, for example, the distribution of the relaxation amount and the evaluation value in the C4 relaxation pattern. The relaxation amount 1401 is the difference between the required amount and the production amount in a plan in which the production amount x _{10Bm is smaller than the required amount with respect to the required amount D 10B of C4.} The evaluation value 1402 is an evaluation value of the target plan using the calculation function of the evaluation index generation unit 203 of the planning unit 103.

FIG. 15 shows, for example, the application frequency of the relaxation amount in the C4 relaxation pattern. The relaxation amount 1501 is the same value as 1401 in FIG. The application frequency 1502 is the number of target plans. Here, the evaluation value distribution and the relaxation amount application frequency of the relaxation pattern of C4 are shown, but in the case of the relaxation pattern of C4 and C2, the evaluation value distribution and the relaxation amount application frequency are similarly calculated in a two-dimensional table. Is also good.

Next, in step S1104 of FIG. 11, the relaxation amount application unit 114 selects some relaxation amount candidates from the distribution of the evaluation function, calculates the distance between the relaxation amount candidates and the application frequency, and plans the candidates having a short distance. It is sent to the unit 103, and it is determined that the constraint condition is satisfied. If the constraints cannot be met, try planning with the next mitigation candidate. If the constraint conditions cannot be satisfied, the relaxation amount of the corresponding relaxation pattern is used.

As a method of selecting some relaxation quantity candidates from the evaluation value distribution, for example, the first derivative of the evaluation value distribution is calculated and the deceleration is obtained. The amount of relaxation with severe deceleration is a candidate. Further, if the constraint condition cannot be satisfied even if the drafting is tried with a plurality of relaxation amount candidates of one relaxation pattern, the relaxation implementation of the corresponding relaxation pattern may be stopped depending on the number of implementations and the implementation time.

FIG. 16 is a diagram showing a screen displayed on the user terminal 105 based on the display data output from the planning unit 103 in this embodiment. In FIG. 16, 1601 is a relaxation pattern display screen, 1602 is a relaxation information display screen, and 1603 is a planning result display screen.

The relaxation pattern display screen 1601 is a screen for displaying related constraints, constraint priorities, and relaxation patterns extracted by the constraint relaxation unit via the planning unit 103.

The mitigation information display screen 1602 is a screen showing the distribution of the mitigation amount and the evaluation value and the mitigation amount application frequency for each mitigation pattern. The distribution of the relaxation amount and the evaluation value is a drawing result based on the result of FIG. The relaxation amount application frequency is a drawing result based on the result of FIG. In addition, the relaxation amount and the evaluation value output from the constraint relaxation unit 104 are also shown.

The planning result display screen 1603 is a screen for displaying the corresponding planning result for each mitigation pattern. As shown in FIG. 16, when all the constraints can be satisfied, the planning result is displayed on the planning result display screen 1603.

As described above, in this embodiment, the priority of the constraint condition is determined from the plan history data and set. Then, for the violation constraint, the related constraint is extracted, the priority is taken into consideration, and the mitigation pattern is generated. Then, for each mitigation pattern, the relationship between the mitigation amount and the evaluation function is extracted from the planning history data, and the application frequency of the mitigation amount is extracted from the planning history data. In addition, for each mitigation pattern, the mitigation amount is determined by referring to the relationship between the mitigation amount and the evaluation function and the application frequency of the mitigation amount, a mitigation plan is formulated, and the evaluation score is good for the drafting result for each mitigation pattern. Adopt the result.

As described above, according to this embodiment, a planning system capable of formulating a well-evaluated plan, a planning device used for the planning system, and a planning method are provided in consideration of the relationship between the amount of relaxation of the constraint and the evaluation value of the plan. Can be provided.

Although the examples of the present invention have been described above, the present invention is not limited to the above-mentioned examples, and includes various modifications. For example, the above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to the one including all the described configurations. Further, although each of the above configurations, functions, processing units, etc. has been described as being realized by software by the CPU interpreting and executing a program that realizes each function, a part or all of them are integrated, for example. It may be realized by hardware by designing with a circuit or the like.

100: Planning device, 101: Order information DB, 102: Equipment information DB, 103: Planning unit, 104: Constraint relaxation unit, 105: User terminal, 106: Planning result DB, 111: Constraint relevance extraction unit, 112: Relaxation pattern classification unit, 113: Relaxation information extraction unit, 114: Relaxation amount application unit, 201: Determinant variable generation unit, 202: Constraint condition generation unit, 203: Evaluation index generation unit, 204: Planning unit, 205: Planning result output unit, 701: CPU (processing device), 702: memory (storage device), 703: interface, 704: network interface, 705: keyboard, 706: screen, 707: mouse, 708: hard disk, 1601: relaxation pattern Display screen, 1602: Relaxation information display screen, 1603: Planning result display screen

Claims (6)

It is a planning device that has a processing device and a storage device and formulates a production plan.
The processing device is
A planning department that creates a plan based on the constraints and evaluation functions generated from the input order information and equipment information of the planned schedule,
When the above constraint condition is violated, it has a constraint relaxation unit that relaxes the violation constraint.
The constraint relaxation unit identifies the constraint related to the violation constraint, generates a relaxation pattern, and for each relaxation pattern, the relationship between the relaxation amount and the evaluation value in the evaluation function from the historical data, and the relaxation amount. The relaxation amount is extracted, and the relaxation amount is selected based on the relationship between the relaxation amount and the evaluation value and the application frequency of the relaxation amount.
The planning unit is a planning device characterized by formulating a plan based on the mitigation amount for each mitigation pattern generated, evaluating with the evaluation function, and adopting a plan having a better evaluation. The planning device according to claim 1.
The constraint relaxation unit is a planning device characterized in that a constraint related to the violation constraint is specified and the relaxation pattern is generated from a combination of the related constraints. The planning device according to claim 1.
The constraint relaxation unit extracts the distribution of the relaxation amount and the evaluation value by the evaluation function from the history data for each relaxation pattern.
A relaxation amount candidate is selected based on the amount of change in the distribution of the evaluation values.
Extract the application frequency of the relaxation amount and
A planning apparatus characterized in that a relaxation amount is selected based on the relationship between the relaxation amount candidate and the evaluation value and the application frequency of the relaxation amount. The planning device according to claim 1.
The planning unit is a planning device characterized by outputting the relaxation pattern, the relationship between the relaxation amount and the evaluation value, the application frequency of the relaxation amount, and the display data of the result of the planning. It is a planning method to make a production plan for inputting order information and equipment information of the planned schedule.
Create a plan based on the order information of the planned schedule and the constraints and evaluation functions generated from the equipment information.
When the constraint condition is violated, the violation constraint is relaxed, the constraint related to the violation constraint is specified, a relaxation pattern is generated, and for each relaxation pattern, the relaxation amount and the evaluation value by the evaluation function are obtained from the historical data. And the application frequency of the relaxation amount are extracted, and the relaxation amount is selected based on the relationship between the relaxation amount and the evaluation value and the application frequency of the relaxation amount.
A planning method characterized by formulating a plan based on the mitigation amount for each mitigation pattern generated, evaluating with the evaluation function, and adopting a plan with a better evaluation. It is a planning system equipped with a planning device that has a processing device, a storage device, and a network interface to formulate a production plan, an order information database that includes order information of the planned schedule, and an equipment information database that includes equipment information.
The processing device is
A planning unit that creates a plan based on the constraint conditions and evaluation functions generated from the order information of the planned schedule input from the order information database and the equipment information input from the equipment information database via the network interface.
When the above constraint condition is violated, it has a constraint relaxation unit that relaxes the violation constraint.
The constraint relaxation unit identifies the constraint related to the violation constraint, generates a relaxation pattern, and for each relaxation pattern, the relationship between the relaxation amount and the evaluation value in the evaluation function from the historical data, and the relaxation amount. The relaxation amount is extracted, and the relaxation amount is selected based on the relationship between the relaxation amount and the evaluation value and the application frequency of the relaxation amount.
The planning unit is a planning system characterized by formulating a plan based on the mitigation amount for each mitigation pattern generated, evaluating with the evaluation function, and adopting a plan with a better evaluation.

Images