Flexible job shop batch dynamic scheduling optimization method

Abstract

The invention discloses a flexible job shop batch dynamic scheduling optimization method. According to the method, a production mode of batch scheduling and the dynamic factor of machine malfunction are introduced into a flexible job shop scheduling problem; relevant constraints are extended as needed with delivery period constraints considered, a BDFJSP problem model adopting minimizing delay time as an objective is established; on the basis of the problem model, a Q learning mode is adopted to train an agent, so that a solution can be obtained, and a reasonable state set, an action set and areward function are established according to the problem; and in order to better balance exploration and utilization in the Q learning, epsilon greedy search is integrated into the Q learning, and areasonable epsilon attenuation function is set. The method is of great significance and engineering practical application value for solving the job scheduling of flexible shops and even discrete optimization problems similar to the job scheduling of the flexible shops.

Description

technical field [0001] The invention relates to the technical field of intelligent optimization of discrete combination problems under dynamic working conditions represented by the technical field of flexible job shop scheduling optimization. Background technique [0002] The flexible job shop scheduling problem is one of the most difficult problems in the manufacturing system. Its complex combination of processing machines and workpieces leads to high computational complexity and is a very strong NP-hard problem. In order to optimize within a reasonable time range, scholars are more inclined to use meta-heuristic intelligent algorithms to obtain approximate solutions (satisfactory solutions) to the optimal solution, thus developing and improving many intelligent algorithms, but most of these algorithms are For static FJSP issues. Many unpredictable dynamic factors often occur in the actual process, such as machine failure, arrival of new workpieces, etc., so the above algo...

AI Technical Summary

Problems solved by technology

At the same time, through our research, the existing literature to solve the dynamic scheduling problem of the job shop all adopts reactive priority rules or meta-heuristic algorithms, which may have insufficient flexibility / intelligence (cannot intelligently deal with different working conditions) or Disadvantages such as requiring a large number of iterative optimizations (leading to reduced efficiency and not meeting the response time requirements of dynamic scheduling)

Images