A Study On Job Shop Scheduling Based On Machine Learning Techniques

As the manufacturing industry tends toward product personalization and diversified manufacturing processes,the job shop scheduling problem has become more complicated.As a classic combinatorial optimization problem,the scheduling problem can effectively solve the problem of optimal allocation of resources.However,in the face of a complex and changeable working environment,traditional production scheduling model optimization and optimization algorithms have encountered bottlenecks.In response to the above problems,this thesis studies the Job Shop scheduling algorithm of decision tree and random forest algorithm based on machine learning thesis.In order to better solve the scheduling problem of the flexible job shop,the characteristics and main performance indicators of the job shop scheduling problem are analyzed,and the mathematical model of the flexible job shop scheduling problem is established;the data characteristics and preprocessing methods of the job shop are studied.The classification and comparative analysis of the dynamic disturbance situation lays the foundation for the follow-up research on the Job Shop scheduling problem based on the decision tree algorithm and the random forest algorithm.Aiming at the flexible job shop scheduling problem,a scheduling algorithm based on decision tree and random forest Job Shop is proposed.Mining new decision trees and random forest rules through job shop scheduling data,and scheduling job shop tasks according to the new rules.Taking a stand-alone system as an example,the simulation and comparison of the EDD rules,the decision tree algorithm and the stand-alone scheduling model of the random forest algorithm are carried out.By comparing the maximum weighted delay time,it can be found that the maximum weighted delay time obtained by the random forest algorithm is the smallest,indicating that the random forest algorithm is correct.Solving job shop scheduling problems has a relatively good effect.Considering the dynamic disturbance of the Job Shop scheduling problem,the rescheduling problem under the disturbance is analyzed.In order to solve the problem of dynamic job shop scheduling under disturbance,according to the internal relationship of data in job shop scheduling,the optimization algorithm of decision tree and random forest is adopted to design a scheduling framework and mine new decision rules.Combining the simulations of two flexible job shops,and comparing the maximum completion time of SPT rules,decision tree algorithm and random forest algorithm through simulation,it is found that the maximum completion time of the flexible job shop simulation example under two dynamic disturbances by the random forest algorithm Both are the smallest,indicating the feasibility and effectiveness of the machine learning-based scheduling algorithm proposed in this thesis.The random forest algorithm is used to provide a feasible scheduling scheme for the scheduling job shop in a complex environment,which provides a reference for application in actual production.