WO2020164293A1

WO2020164293A1 - Method and system for intelligently generating production scheduling of enterprise

Info

Publication number: WO2020164293A1
Application number: PCT/CN2019/120979
Authority: WO
Inventors: 刘煜; 孙再连; 梅瑜
Original assignee: 厦门邑通软件科技有限公司
Priority date: 2019-02-11
Filing date: 2019-11-26
Publication date: 2020-08-20
Also published as: CN109948899A

Abstract

A method and system for intelligently generating a production scheduling for an enterprise, wherein the described scheduling method and system, by means of a human-machine chess game, automatically generate a new production scheduling according to changes to production scheduling factors. The described new production scheduling uses a production scheduling assessment indicator as a target, uses meeting production requirements as a precondition, uses changes to production scheduling factors as a constraint condition for playing chess, and solves the problem of production scheduling for an enterprise by using the thinking of a human-machine chess game so as to change the complex into the simple; the described new production scheduling considers unpredictable changes to production scheduling factors to be constraint conditions for playing chess, i.e., considers same to be the moves of an opponent during a human-machine chess game, thereby changing complex change factors into routine actions of an opponent; and the scheduling system performs a next step reaction according to the action of an opponent to provide a new production scheduling; meanwhile, a production scheduling suggestion of the new production scheduling is a historical scheduling that originates from historical knowledge, that meets a constraint condition, and that has an optimal comprehensive indicator, and the scheduling suggestion provided thereby is secure, reasonable, optimal and reversible.

Description

A smart generation enterprise production scheduling method and system

Technical field

The present invention relates to the technical field of production management, in particular to a production scheduling method and system for smart generation enterprises.

Background technique

The current production scheduling plan of various industries is basically the APS system for monthly production scheduling.

technical problem

However, it is still often necessary for engineers to adjust the production plan of production equipment based on experience, especially the frequent changes in scheduling factors such as temporary insertion, plan changes, untimely arrival of auxiliary materials, insufficient raw materials, etc., all require experienced and dedicated responsibility If the master is not present, it will be difficult for others to deal with it. The above-mentioned traditional production management methods have low efficiency, high labor costs, and poor stability, which cannot meet market demands.

Technical solutions

In order to solve the above technical problems, the present invention provides a smart generation enterprise production scheduling method and system, wherein the method and the scheduling system adopt artificial intelligence thinking and technology in a man-machine game mode, which can affect people who affect production scheduling. Analyze the five major links of machine, material, and law, and automatically generate new production schedules according to changes in scheduling factors.

The new production schedule takes the production scheduling evaluation index as the goal, meets the production demand as the premise, takes the change of production scheduling factor as the constraint condition of the game, and adopts the thinking of man-machine game to solve the problem of enterprise production scheduling, reducing complexity to simplicity .

Optionally, in the human-machine game mode, the production machine is regarded as a finite state machine or chess piece, and the M types of products that the production machine can produce are regarded as the M states or M states of the finite state machine or chess piece. In this way, the scheduling system can list all feasible routes in advance according to the changes in scheduling factors and the way the production machine moves.

Optionally, the scheduling system learns the production product type, production capacity, and product switching route of each production machine, and transforms it into alternative directions and continuous walking routes for each finite state machine or chess piece .

Optionally, the constraint conditions include changes to the original production plan, raw material reserves, delivery deadlines, equipment conditions, etc.; the constraint conditions restrict the production machine's selection of products to be produced; that is, under the premise of meeting production requirements Next, the production schedule of the previous day is regarded as the current board layout, and the constraint conditions are regarded as the opponents of the scheduling system. The scheduling system exhaustively plans all routes within N days, and eliminates impossible routes with constraints , Restrict the next walking route of the finite state machine or the chess piece; the next walking route is different, corresponding to a different new production schedule.

Optionally, the scheduling system learns historical scheduling and real-time scheduling, that is, learning historical routes and real-time moves of the finite state machine or chess pieces; the scheduling system exhaustively plans all historical routes within N days, Select the historical route with the best production scheduling evaluation index in the history; when the demand and the constraint conditions are not met in the historical route, the route will be given according to the default scheduling plan, which is selected randomly A commonly used move.

Optionally, the production scheduling evaluation index is a rule for evaluating the pros and cons of the new production scheduling, and the production scheduling evaluation index is obtained by a weighted comprehensive scoring of indicators concerned by the user.

Optionally, the indicators include a rebrand score, a pre-inventory score, and a pre-delivery period score. The scheduling evaluation index=w1* rebrand score+w2*pre-stock score+w3 pre-delivery period score, where w1, w2 and w3 are the weights of each index.

Optionally, the rebrand score: determine whether each production machine has the same brand on the same day as the previous day. If not, record 1; otherwise, record 0. Summarize the number of rebrands of all machines in the planned N days. The total number of cards changed in the new production schedule is arranged in descending order, and the Kth place is scored with K points;

Pre-inventory score: Calculate the output of each brand after the pre-planned production, deduct the pre-allocation amount of the corresponding brand, and judge whether the difference is within the upper and lower limits of the respective inventory. If the difference is within the upper and lower limits of the respective inventory, mark 1 for the brand, otherwise mark 0, and finally summarize each The pre-stock total fulfillment number of the brand is arranged in ascending order, and the K-th place is scored K.

Pre-delivery period score: within N days of the plan, first judge whether the delivery period of each brand is within the N-day plan. If a label is calculated for the brand with the delivery period during this period, add 1 if each brand can be delivered in advance. If the delivery can be completed on time, add 0 to accumulate the counts and arrange them in ascending order, and score K for the Kth place.

Correspondingly, a smart generation enterprise production scheduling system is based on the above-mentioned smart generation enterprise production scheduling method.

Beneficial effect

From the above description of the present invention, it can be seen that compared with the prior art, the present invention proposes a smart generation enterprise production scheduling method and system, which is the first to use human-machine game thinking and technology to solve the problem of enterprise production scheduling. In the scheduling process, the complexity is simplified, and the unpredictable changes in scheduling factors, such as inserting orders, plan changes, and insufficient raw materials arrival, are all regarded as constraints of the game, that is, as a human-machine game. , The opponent’s move, the complex change factor is turned into the opponent’s routine action, the scheduling system will make the next response according to the opponent’s action, give a new production schedule, and at the same time, give a new production schedule, The scheduling recommendations are derived from historical knowledge, satisfy the constraints, and have the best comprehensive indicators. The scheduling recommendations given by it are safe, reasonable, optimal and flexible.

To

Embodiments of the invention

In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention clearer and clearer, the present invention will be further described in detail below in conjunction with embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, but not to limit the present invention.

Embodiment 1: A smart generation enterprise production scheduling method. In the method, the scheduling system adopts artificial intelligence thinking and technology in a human-computer game mode to affect the five major links of human, machine, material, and law that affect scheduling Perform analysis and automatically generate new production schedules based on changes in scheduling factors. The new production schedule takes the production scheduling evaluation index as the goal, meets the production demand as the premise, takes the change of production scheduling factor as the constraint condition of the game, and adopts the thinking of man-machine game to solve the problem of enterprise production scheduling, reducing complexity to simplicity .

Specifically, in the human-machine game mode, the production machine is regarded as a finite state machine or chess piece, and the M types of products that the production machine can produce are regarded as M states or M types of the finite state machine or chess piece. How to go, the scheduling system can list all feasible routes in advance according to changes in scheduling factors and the way of production machines.

Of course, the scheduling system also needs to learn the production capacity of each production machine, product switching routes, and other factors that affect the direction and selection of the finite state machine or chess pieces. In the specific use process, according to the product Type and plan quantity, get schedule schedule.

In this embodiment, the constraint conditions include changes to the original production plan, raw material reserves, delivery deadlines, equipment conditions, product types, pre-allocated amounts, and the number of boxes that can be produced. The constraint condition restricts the selection of the production machine by the production machine; that is, under the premise of meeting the production demand, the production schedule of the previous day is regarded as the current board layout, and the constraint condition is regarded as the schedule For the opponent of the system, the scheduling system exhausts all the routes planned in N days, excludes impossible routes with constraint conditions, restricts the finite state machine or the next walking route of the chess pieces, and gives new production scheduling suggestions. The following different walking routes correspond to different new production schedules.

Embodiment 2: In order to obtain scheduling suggestions with safety, rationality, optimality, and flexibility, this embodiment is based on Embodiment 1, and the scheduling system learns historical scheduling and real-time scheduling , That is, learn the historical routes and real-time moves of the finite state machine or chess pieces; the scheduling system exhaustively planned all the historical routes within N days, and selects the historical route with the best scheduling evaluation index in the history; when the When the requirements and the constraint conditions are not met in the historical route, the walking method is given according to the default scheduling plan, and the default scheduling plan is to randomly select a common walking method.

Embodiment 3: In the first and second embodiments, the scheduling system may provide a variety of scheduling plans for the company to choose from, and the company will choose the best scheduling plan with the highest scheduling evaluation index. In this embodiment, a specific calculation method of scheduling evaluation index is given:

The scheduling evaluation index is a rule for evaluating the pros and cons of a new production scheduling, and the scheduling evaluation index is obtained by a weighted comprehensive scoring of indicators concerned by the user.

The indicators include rebrand score, pre-inventory score, and pre-delivery period score. The scheduling evaluation index = w1* re-brand score+w2*pre-inventory score+w3 pre-delivery period score, where w1, w2, and w3 are The weight of each indicator. In other embodiments, or in different enterprises, the corresponding indicators and indicator weights may be different.

Among them, the re-brand score: judge whether each production machine has the same brand on the day and the previous day. If not, record 1, otherwise, record 0. Summarize the number of license changes of all machines within N days of the plan. The total number of card changes in the production schedule is arranged in descending order, and the Kth place is scored with K points;

The fourth embodiment corresponds to a smart generation enterprise production scheduling system, which is based on the smart generation enterprise production scheduling method of the foregoing embodiment.

In specific applications, the scheduling system needs to obtain basic information and production scheduling preparations of the enterprise, and prepare the scheduling space and learn the remaining plan set before scheduling. The scheduling system also needs to learn historical scheduling, and update the scheduling knowledge base according to actual schedules that occur every day.

Wherein, the basic information specifically includes the following:

1. Machine capacity table: machine name, machine hourly quota, actual daily production time, production mode

Product type: category, brand number sequence; (product type is the common basic attribute of a type of product with the same production mode. For example: Red wolf and gray wolf are both in the same type of production mode, so they are classified as category 1)

2. Product name category table: brand, production mode (each production mode corresponds to Y brand)

] 3. Inventory upper and lower limit table: grade, inventory upper limit, inventory lower limit

4. Hot-selling status of grades: grade, yes or not (this configuration table records the hot-selling status of the brand in the current season, it is recorded as 1, otherwise it is recorded as 0)

5. Constraint configuration table: product type, delivery period, pre-allocation, number of boxes that can be produced, estimated material availability date, estimated material production volume

The production scheduling preparation specifically includes the following:

1. Production plan: Summarize the production plan and production schedule every day, and calculate the remaining production plan.

2. Summary of constraint conditions: production plan, picking plan, inventory rules, and auxiliary material plan are specific to each product name category.

(Provide a method to check whether the current schedule meets the constraints.)

1), production plan:

A. When the inventory is greater than the sum of the pre-allocated amount and the lower limit of the inventory, no production scheduling will be made for this brand;

B. When the inventory is less than the sum of the pre-allocated amount and the lower limit of inventory, priority is given to scheduling if the inventory is less than the pre-allocated amount, followed by the case where the inventory is greater than the pre-allocated amount but less than the sum of the pre-allocated amount and the lower limit of inventory. Produce.

2). Delivery plan:

[0045] Calculate the time difference between the day of the brand and its delivery date. A small time difference indicates that it is urgent to prioritize production. That is, they are arranged in ascending order of time difference, and the higher the ranking, the higher the priority.

3). Inventory rules:

A. The pre-stock quantity cannot be greater than the upper limit of the stock;

B. When the total amount of pre-stock must not exceed the warehouse saturation value.

4), auxiliary material plan:

A. When the auxiliary materials and raw materials are sufficient, priority is given to scheduling.

B. When auxiliary materials and raw materials are insufficient, plan according to how many boxes can be scheduled, and then plan the remaining quantity according to the estimated available date and production capacity.

3. Machine production base

The production base of a machine is the number of boxes that a machine can produce in one day.

According to the remaining production plan, calculate the production base of each product name category.

When the scheduling space is prepared, an X-day scheduling temporary storage area is established, an empty optimal scheduling table is established, and the current most efficient scheduling evaluation index is 0.

During the scheduling process, the specific steps are:

S100: Obtain the latest remaining plan set and the latest constraint conditions;

S200: Traverse each product name category;

S210: Obtain product type;

S220: Traverse the scheduling knowledge base of the product type;

S221: Find historical schedules of the same product type that are less than or equal to the maximum cardinality of the remaining plans, and when the constraint conditions are also suitable for the current situation, schedule the details of the historical schedules into the schedule temporary storage area. If there is no historical schedule that meets the conditions, the default scheduling plan is used, and the default scheduling plan meets the constraint conditions and the remaining plan set, regardless of whether it is the rightmost;

S222: If the historical schedule selected in S221 does not meet the constraints of the product name category, go back to step 2.2 .1, Re-obtain a new historical schedule;

S230: Calculate the scheduling evaluation index, and if it is greater than the optimal scheduling evaluation index, copy the schedule to the optimal scheduling table;

S300: Save the optimal scheduling table in the database of the scheduling system;

S400: End.

The present invention has been described exemplarily above. Obviously, the specific implementation of the present invention is not limited by the above-mentioned manners, as long as various insubstantial improvements made by the method concept and technical solution of the present invention are adopted, or the present invention has not been improved. The concept and technical solution of the invention are directly applied to other occasions, all within the protection scope of the present invention.

To

Claims

A method for intelligently generating enterprise production scheduling, which is characterized in that the scheduling system automatically generates a new production schedule based on changes in scheduling factors in a man-machine game mode; the new production schedule uses a scheduling evaluation index as The goal is to meet the production demand as the premise, and the change in production scheduling factors is the constraint of the game.
The production scheduling method of a smart generation enterprise according to claim 1, characterized in that, in the human-machine game mode, the production machine is regarded as a finite state machine or chess piece, and the production machine can produce M The products are regarded as the M states or M moves of the finite state machine or chess pieces.
The smart generation enterprise production scheduling method according to claim 2, wherein the scheduling system learns the production product type, production capacity, and product switching route of each production machine, and transforms them into each Alternative directions and continuous walking routes of the finite state machine or chess pieces.
The production scheduling method of a smart generation enterprise according to claim 3, wherein the constraint conditions include changes to the original production plan, raw material reserves, delivery deadlines, and equipment status; the constraint conditions restrict the production machine The choice of products to be produced; that is, under the premise of meeting production requirements, the production schedule of the previous day is regarded as the current board layout, and the constraint conditions are regarded as the opponent of the scheduling system, and the scheduling system Exhaust all routes within N days of the exhaustive plan, exclude impossible routes with constraint conditions, and restrict the next walking routes of the finite state machine or chess pieces; the next walking routes are different and correspond to different new production schedules.
The production scheduling method of a smart generation enterprise according to claim 4, wherein the scheduling system learns historical scheduling and real-time scheduling, that is, learning historical routes and real-time moves of the finite state machine or chess pieces. Method; the scheduling system exhaustively planned all historical routes within N days, select the historical route with the best scheduling evaluation index in the history; when the historical route does not meet the requirements and the constraints, press the default The production scheduling plan provides a walking method, and the default production scheduling plan is to randomly select a commonly used walking method.
The method for intelligently generating enterprise production scheduling according to any one of claims 1 to 5, wherein the scheduling evaluation index is a rule for evaluating the pros and cons of a new production scheduling, and the scheduling evaluation index Obtained by the weighted comprehensive score of the indicators concerned by users.
A production scheduling system for a smart generation enterprise, characterized in that the scheduling system is based on the production scheduling method for a smart generation enterprise according to any one of claims 1 to 6.