JPH1179408A - Raw material yard operation plan forming method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a raw material yard operation plan forming method whereby a practical operation restriction rule can be perfectly reflected to a plan, enhance forming accuracy for the practical plan. SOLUTION: In order to properly perform work receiving and delivering raw material yard temporarily storing a raw material and properly arrange it, a necessary data is input. A decision logic method is applied, a search range of solution related to acceptance, delivery, arrangement, etc., of the raw material is limited. With the search range of solution thus limited as an object, genetic algorithm is applied, an optimum solution or semi-optimum solution is searched. Based on a raw yard operation data including the searched optimum solution or semi-optimum solution, the optimum plan is planned.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for preparing an optimal schedule of an operation schedule for receiving and discharging raw materials in a raw material yard for temporarily storing raw materials such as iron ore and coal in a steel mill. It is.

[0002]

2. Description of the Related Art In a yard for temporarily storing raw materials such as iron ore and coal at an ironworks, the receiving and dispensing of raw materials is carried out in order to secure an empty space for receiving raw materials within a limited area. This is performed by a method in which an operator determines, based on experience, a vacant lot for receiving raw materials and a mountain for dispensing raw materials, while considering various operation rules and operational constraints. However, in these methods, it is difficult to quantitatively evaluate the determination result because the determination is made by a skilled operator, and it is difficult to improve the planning accuracy because the result depends on the individual skill.

Further, a method of applying a genetic algorithm has been reported as a method of preparing an operation plan for a raw material yard. For example, there is a “method of creating a yard operation plan for a raw material yard” proposed in Japanese Patent Application Laid-Open No. 8-40560.
This planning method applies a genetic algorithm to a population that uses the receiving information and the payout information within the planning period as genetic information, and performs crossover and mutation, as well as selection, propagation and selection based on a predetermined fitness function. It is to obtain the optimal plan repeatedly. In addition, rules that can be expressed as a probability function among the operation rules can be incorporated in this plan creation method.

[0004] This plan creation method is a method for creating an optimal plan only by using a genetic algorithm, and thus cannot fully incorporate absolute operation constraints.
Although the created solution has a high evaluation under a certain evaluation function, there is a problem that it does not conform to practical operation constraint rules.

[0005] For example, an operation of receiving raw materials of the same brand so as to cover a pile of raw materials existing as stock in a raw material yard (hereinafter referred to as Motoyama cover), or receiving the entire amount of the received brand in one open space Can not,
The operation of dividing into several vacant lots and accepting the raw materials when it should not be done (hereinafter referred to as lot division) cannot be expressed. Therefore, the created solution does not satisfy the operation constraints such as Motoyama cover and lot division,
It is necessary to measure the improvement of the created solution by some correction means, and further, it may occur that the solution cannot be corrected depending on the content of the created solution.

[0006]

In a method of preparing an operation plan of a raw material yard using only a genetic algorithm, information on the reception of raw materials within a plan period (reception date, acceptable vacant land management number, reception direction) ) And payout information (payment date, payout brand pile management number, payout direction). However, according to this method, as described above, the operation constraint rules cannot be fully incorporated into the plan,
Although the created solution is a highly evaluated solution under a certain evaluation function, there is a problem that it does not conform to practical operational constraint rules.Therefore, it is necessary to improve the created solution by some means. Occurs.

[0007] It is therefore an object of the present invention to be able to fully reflect the practical operational constraint rules in the plan, and
An object of the present invention is to provide a method for preparing a raw material yard operation plan with high accuracy in preparing a practical plan.

[0008]

Means for Solving the Problems In order to achieve the above object, the inventor has conducted intensive studies. As a result, in preparing the operation plan for the raw material yard, a part of the plan that satisfies the operation constraint rules was created by a deterministic method, for example, a method using artificial intelligence, and the search space for the solution was limited to some extent. Then, it has been found that by applying a genetic algorithm to the search space of such a limited solution and creating a remaining plan, a raw material yard operation plan with high creation accuracy can be created.

The method for preparing a raw material yard operation plan according to the present invention comprises:
It is based on the above-mentioned knowledge, and performs the following steps for appropriately performing the receiving and dispensing work of the raw material yard for temporarily storing the raw material, and for appropriately arranging the raw material. Characterized by the step of inputting necessary data, the step of applying deterministic methods to limit the search range of the solution with respect to the acceptance, dispensing and arrangement of raw materials, in this way Applying a genetic algorithm to the limited search range of the solution to search for an optimal solution or a sub-optimal solution, and based on the raw material yard operation data including the searched optimal or sub-optimal solution. The process of developing an optimal plan.

Further, in the method for preparing a raw material yard operation plan according to the present invention, the step of limiting the search range of the solution includes a process of reducing the number of solution variables by adding an operational constraint. It is assumed that.

Further, in the method for preparing a raw material yard operation plan according to the present invention, the step of applying a genetic algorithm includes the step of applying the genetic algorithm to a group of individuals having the receiving information and the paying-out information within the period as genetic information. , Mutation, and repeated selection, multiplication and selection based on a predetermined fitness function.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A system used in a method for preparing a raw material yard operation plan according to the present invention includes a raw material yard process control having data necessary for preparing a raw material yard operation plan and an engineering for preparing the plan. It consists of a workstation (EWS).

In the step of inputting necessary data,
Data such as stock status of each brand in the raw material yard, storage status of each brand in the raw material yard, information on receiving raw materials into the raw material yard, information on dispensing raw materials from the raw material yard, and the like are input.

In the step of applying a deterministic method to limit the search range for solutions relating to the acceptance, dispensing, arrangement, etc. of raw materials, an operation restriction rule (operation restriction is applied) by an artificial intelligence expert system or the like. In addition, a part of the solution based on (reducing the number of solution variables) is searched and determined as a part of the raw material yard operation plan, and the search range (space) of the solution is limited to some extent.

In the step of searching for an optimal solution or a sub-optimal solution by applying a genetic algorithm, an individual whose acceptance information and payout information within a period are genetic information is targeted for a limited search range of the solution. Applying a genetic algorithm to the group, repeating crossover, mutation, and selection, multiplication and selection based on a predetermined fitness function,
Search for an optimal or suboptimal solution. The method of expressing an individual and the method of genetic manipulation used when applying the genetic algorithm may be those conventionally used.

In the step of drafting an optimal plan, an optimal plan is drafted based on the raw material yard operation data including the searched optimal solution or suboptimal solution. According to the method for preparing a raw material yard operation plan of the present invention, as described above, in preparing a raw material yard operation plan, one of the plans that satisfies the operation restriction rules by a deterministic method, for example, a method using artificial intelligence. A part can be created to limit the search space for the solution to some extent, and the remaining plan can be created by applying a genetic algorithm to the solution space limited in this way. Further, as described above, the practical operation constraint rules can be completely reflected in the plan, and the raw material yard operation plan with high accuracy in preparing the practical plan can be created.

[0017]

BRIEF DESCRIPTION OF THE DRAWINGS An embodiment of the method according to the invention is described below with reference to the drawings. FIG. 1 is a diagram showing one embodiment of the system configuration of the method of the present invention. The system of the method of the present invention comprises a raw material yard process control having data necessary for preparing a raw material yard operation plan and an engineering workstation (EW) for preparing a plan.
S). The raw material yard process computer has a database (DB) storing the stock status of each brand in the raw material yard, a DB storing the storage status of each brand in the raw material yard, and information on receiving raw materials into the raw material yard. And a DB storing raw material payout information from the raw material yard, and a function of creating a raw material receiving / paying instruction.

EWS creates a raw material yard operation plan based on each DB information of the raw material yard process control, and
It has the function of displaying. In the deterministic method processing section,
For example, a part of the solution based on the operation restriction rule is searched for by the expert system of artificial intelligence and determined as a part of the raw material yard operation plan, and the search space for the solution is limited to some extent. The GA processing unit applies a genetic algorithm to the limited solution space as described above to create a remaining plan.

The final raw material yard operation plan created by the GA processing unit is drawn by the raw material yard operation plan diagram creation function unit and is guided to the operator. The raw material yard operation plan confirmed by the operator is transmitted to the raw material yard process control, and specific receiving instructions such as a receiving yard number, a position in the yard, a brand name, a receiving amount, and a receiving time are sent to the receiving / dispensing instruction preparing function section. And payout instructions such as payout yard number, yard position, brand name,
A payout amount, a payout time, and the like are created.

FIG. 2 is a diagram showing a gene structure used in the genetic algorithm. In FIG. 2, daily acceptance plan information and payout plan information are arranged in the order of the planned date. The locus for which the acceptance information has been created incorporates information on the name of the accepting brand, the vacant lot to accept, and the accepting direction. At the locus where the payout information is created, information on the name of the payout brand, the peak to be paid out, and the payout direction is incorporated.

FIG. 3A is a flowchart showing a procedure for preparing a raw material yard operation plan. The outline is as follows. (1) First, various related data are input. For example,
Each DB data of the raw material yard process control shown in FIG. 1 corresponds thereto. (2) Next, the search space for the solution is limited to some extent by a deterministic method. For example, there are restrictions such as brands that need to be covered by Motoyama, brands that need to be divided into lots, brands that need to specify receiving vacant lots / directions, or brands that need to specify a payout pile / direction. In addition, a process of reducing the number of solution variables is performed. (3) Next, using the genetic information shown in FIG. 2, a genetic algorithm is applied to the solution search space limited to some extent in the above-described processing to search for an optimal solution or a sub-optimal solution. (4) Next, a yard operation plan diagram is drawn from the final raw material yard operation plan data (that is, the optimal solution or the sub-optimal solution) created in the above item (3).

FIG. 3B is a diagram showing in detail how to create a plan for applying the genetic algorithm of FIG. 3A. The outline is as follows. (1) First, an appropriate number of solution candidates that can be represented by the gene structure shown in FIG. 2 are created, and each of them is referred to as an individual, and an individual group of the first generation is created. (2) Next, to make the number of the first generation one generation,
It is assumed that the generation number N = 1. (3) Next, the individuals are crossed over each other.
Specifically, the daily payout and acceptance information are exchanged with each other. (4) Next, a certain individual is mutated. Specifically, information on a certain locus, for example, brand,
Change the information such as the payout mountain. (5) Next, individuals to be left in the next generation are selected based on the value of the evaluation function. Specifically, individuals with high evaluation values are selected.

Evaluation function f = α1Σ (vacant land area) ³ + α
2Σ (moving time of mobile station) α1, α2: weight (6) Next, in order to set the next generation number to N + 1, the generation number N = N + 1. (7) Next, an Nth generation population is created by the processing of the above items (3) to (6). (8) Next, it is determined whether N is equal to or greater than a set value. N
Is smaller than the set value, the process returns to the process of the above item (3). N
If is equal to or greater than the set value, the process ends.

Next, an embodiment of the method of the present invention will be described in a very simple case where the planning period is 30 days. In this simple case, the number of raw material yards is 1, α1 =
1, α2 = 0. (1) The initial yard map is the state shown in the upper part of FIG.

Yard length: 1000 m Brand A is placed at a position of 0 to 100 m in the yard, 10,000 tons or less, B 200 to 400 m 20,000 tons C 500 to 600 m 10,000 tons D 700 to 1000 m 30,000 tons ing. (2) Receiving information Receiving date Brand Accepted amount Restriction rule 5th day A 10,000 tons Motoyama covering 15th day E 25,000 tons Lot division 25th day F 20,000 tons None (3) Payment information Payment period Brand Amount to be paid out per day 1 to 20 days B 1,000 tons 6 to 30 days D 800 tons 11 to 30 days A 500 tons 11 to 20 days C 1,000 tons 21 to 30 days E 1,000 FIG. 4 is a yard operation diagram for 30 days in which an initial yard map is described in the upper part. In FIG. 4, the decrease of the mountain in the yard and the addition of a new mountain are shown. In the initial state of the yard, raw materials of brand A are loaded from 0 to 100, vacant lots from 100 to 200, and 200 to 4
Raw material of brand B is loaded until 00, empty lots are loaded from 400 to 500, raw material of brand C is loaded from 500 to 600, empty lots are loaded from 600 to 700, 700 to 1000
Until then, the raw material of Brand D is loaded.

A daily operation plan is shown below the initial yard map. When paying out Brand B, Brand B
The width of the mountain gradually narrows downward. When accepting the accepted brand F, a peak of the brand F is formed in a place that was vacant on the 25th day. The acceptance of the issue A with the restriction rule of covering the original mountain is formed so that the width of the peak of the issue A is widened on the fifth day. When accepting the brand E with the restriction rule of lot division, two mountains of the brand E are formed in a place that was vacant on the 15th day.

As a result of applying the planning method of the present invention and repeating the calculation up to an appropriate generation N, one day after 30 days,
Aggregated to 00-400m and 880-1000m,
Maximization of wide open areas was obtained.

In the present embodiment, the number of yard surfaces is one, but the present invention can be applied to a plurality of yard surfaces.

[0029]

According to the method for preparing a raw material yard operation plan of the present invention, a certain evaluation function is appropriately set for the acceptance and withdrawal information and the initial yard map of all brands within a certain period, thereby restricting the operation. An optimal yard operation plan that adheres to the rules can be created, the accuracy of the plan creation has been greatly improved, and the time taken to create the plan has been significantly reduced.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram of a method of the present invention.

FIG. 2 is a diagram showing a gene structure in the method of the present invention.

FIG. 3 is a flowchart showing a procedure for creating the method of the present invention.

FIG. 4 is a yard operation plan diagram of a simple case according to an embodiment of the method of the present invention.

Claims (3)

[Claims] 1. A raw material yard, which comprises the following steps for properly performing a receiving and discharging operation of a raw material yard for temporarily storing raw materials and for appropriately arranging the raw materials. Operation plan creation method: a step of inputting necessary data, a step of applying a deterministic method, and limiting a search range of a solution regarding acceptance, dispensation, arrangement, etc. of raw materials, Applying a genetic algorithm to a solution search range to search for an optimal solution or a sub-optimal solution, and formulating an optimal plan based on the raw material yard operation data including the searched optimal or sub-optimal solution Process. 2. The raw material yard according to claim 1, wherein said step of limiting the search range of the solution comprises a process of reducing the number of variables of the solution by adding operational constraints. How to create an operation plan. 3. The step of applying a genetic algorithm comprises applying the genetic algorithm to a population having genetic information based on the acceptance information and the payout information within a period,
The method for preparing a raw material yard operation plan according to claim 1 or 2, characterized by repeating selection, multiplication and selection based on a mutation and a predetermined fitness function.