Research On Intelligent Scheduling Optimization Algorithm Of Flexible Job-shop

As a well-known non-deterministic polynomial problem,the flexible job-shop scheduling problem is one of the most critical problems in manufacturing and process planning.The purpose of this paper is to study the flexible job-shop scheduling problem to minimize the maximum completion time.At present,many intelligent algorithms have been applied to solve such problems,but the key parameters of most of the improved intelligent algorithms cannot be dynamically adjusted during the solution process,resulting in the solution results that cannot better meet the needs of production.In addition,the current real flexible job-shops face frequent insertion of new jobs.Enterprises will manufacture machines that are more in line with actual production according to the functional requirements of customers.In the job-shop,new jobs will be temporarily processed due to the needs of users,which leads to the problem of dynamic flexible job-shop scheduling problem.Therefore,by rationally arranging the processing sequence of the job and the relationship between the process and the machine,the final completion time can be shortened,and the economic benefit and the utilization rate of the processing machine can be improved.The main contents and conclusions of this study are as follows:（1）Modeling of job-shop scheduling problem.Firstly,the flexible job-shop scheduling problem is described.Secondly,various constraints and mathematical expressions of the scheduling problem are analyzed.Then the objective function is determined with the shortest makespan as the goal.Finally,the coding method corresponding to the operation sequence and the machine allocation is set up,and the modeling of the flexible job shop scheduling problem is completed.（2）Design of job-shop scheduling algorithm.First,the principles of artificial bee colony algorithm and Q-learning algorithm are introduced.Secondly,combining the two basic algorithms,a self-learning artificial bee colony algorithm is proposed,which uses the exploration and utilization characteristics of the Q-learning algorithm to dynamically adjust the update dimension of each iteration of the artificial bee colony algorithm,and improves the convergence speed and accuracy of the artificial bee colony algorithm.Then,the state,action,reward and action selection strategy in the running of the self-learning artificial bee colony algorithm are set,and the process of the self-learning artificial bee colony algorithm is introduced.Finally,the dynamic scheduling method is determined,and a dynamic self-learning artificial bee colony algorithm is proposed,that is,when a new job is inserted at a certain moment,the operation that has not started processing and the new job are re-scheduled to obtain a new scheduling scheme and reduce the makespan of the job-shop.（3）Job-shop scheduling algorithm verification.First,the self-learning artificial bee colony algorithm is compared with the traditional artificial bee colony algorithm and other improved artificial bee colony algorithms,which proves that the accuracy and speed of the self-learning artificial bee colony algorithm are improved.Secondly,the self-learning artificial bee colony algorithm is compared with other optimization algorithms,including genetic algorithm,gray wolf algorithm,particle swarm algorithm and tabu search algorithm,which further proves the advantages of the algorithm.Then,a dynamic scheduling experiment is designed to verify the effectiveness of the dynamic self-learning artificial bee colony algorithm.Finally,the self-learning artificial bee colony algorithm is applied to the bar production and bundling job-shop,and a scheduling software platform is developed to prove the practicability of the algorithm.