Multi-agent-based job-shop multi-objective collaborative optimization scheduling method

Abstract

The invention discloses a multi-agent-based job-shop multi-objective collaborative optimization scheduling method, which comprises the following steps: a multi-agent-based production workshop scheduling model is established; a workpiece agent proposes a job scheduling request to a planning agent according to a processing task of itself; the planning agent inserts a time cycle window, and carries out multi-objective optimization scheduling based on time, load, cost and energy consumption on a plurality of job scheduling requests received in the window; and the planning agent releases a scheduling result to a device agent, and the device agent finishes the processing task and informs the workpiece agent to release a next process. By inserting the time window, arranging the planning agent and packaging a multi-objective optimization algorithm, optimization range is expanded, overall scheduling performance of the system is improved, multi-objective optimization scheduling is realized, and the defects of single-step optimization and single-objective optimization of a conventional multi-agent method are solved.

Description

technical field [0001] The invention relates to the technical field of workshop scheduling control, in particular to a multi-agent-based multi-objective collaborative optimization scheduling method for a workshop. Background technique [0002] As the core technology of advanced manufacturing technology and modern management technology, effective workshop scheduling is a necessary measure to solve the diversification of orders and complex information in today's manufacturing environment. Therefore, how to establish a distributed, open, and efficient intelligent scheduling method, in It is especially necessary in today's manufacturing environment. The development of Multi-Agent System (MAS) provides an effective solution to the above problems. However, the traditional multi-agent method for solving the Job-Shop Problem (JSP) has the following problems: (1) The traditional multi-agent method for solving the job-shop dynamic scheduling problem generally takes a single completio...

AI Technical Summary

Problems solved by technology

However, the traditional multi-agent method for solving the Job-Shop Problem (JSP) has the following problems: (1) The traditional multi-agent method for solving the job-shop dynamic scheduling problem generally takes a single completion time as the optimization goal , in the actual scheduling, not only the time index, but also the production index such as energy consumption and cost should be considered comprehensively. Multi-objectiveness is a basic attribute of the workshop scheduling problem. (2) When the traditional multi-agent method solves the job shop scheduling problem, it often introduces the bidding mechanism in the market as a negotiation strategy between agents, thus causing a single agent to behave as much as possible. It is possible to complete the task goals set by itself while ignoring the collaborative optimization relationship between each other's tasks, which is reflected in the "selfishness" and "selfishness" of a single agent, and there is no "overall view"; (3) traditional multi-intelligence The agent method realizes the task allocation and optimization between agents through the bidding mechanism. However, the task optimization allocation between agents under this mechanism is essentially a single-step, local optimization, and the optimization interval is too small, which leads to the overall scheduling of the system. Performance improvements are very limited

All in all, the current multi-agent method for solving the workshop scheduling problem has the problems of single optimization goal, no global view of a single agent, and too small optimization interval, which causes a gap between theoretical research and practical application

Images