Research On Flexible Job Shop Rescheduling Based On Machine Learning

The rapid development of society has driven the diversified and personalized product demand of the public,and at the same time has a great impact on the production environment of enterprises,making the uncertain factors in the production environment sudden increase,bringing great challenges to the production of enterprises.In order to better meet the needs of the market and the public,at the same time to achieve automation,intelligent production standards,flexible manufacturing production mode has become the best solution in recent years,the extremely important part of this production mode is production scheduling.In order to adapt to the changing production environment and market environment,and break the shackle that the traditional shop is difficult to adjust at any time according to the actual situation,dynamic scheduling or rescheduling has become an important technical means of flexible job shop scheduling problem.In view of the changeable production environment,it is necessary to establish an effective disturbance monitoring mechanism or rescheduling trigger mechanism to intervene at a reasonable time,so as not to affect the production schedule and reduce enterprise benefits.The main means of intervention is to adjust or rearrange the production scheme through rescheduling,with the purpose of reducing the impact of disturbance and ensuring the normal production of the enterprise.Therefore,this paper analyzes the existing rescheduling theory,and on this basis,studies the rescheduling trigger and rescheduling method of flexible job shop after disturbance occurs(namely,when and how to rescheduling),and systematically studies the rescheduling problem of flexible job shop.First of all,the types and characteristics of common disturbance factors in the manufacturing shop are described in detail,and the disturbance category is divided according to the disturbance degree as the definition standard.In order to solve the problem of uniform quantization of all disturbance factors,a disturbance quantization method based on disturbance delay was proposed,so that all kinds of disturbances could be quantized and compared uniformly,which laid a good foundation for the subsequent research.Secondly,each process in the flexible job shop is analyzed in detail,and the dynamic correlation between the processes is summarized.According to this correlation,the rescheduling moment point theory is proposed to standardize the rescheduling trigger time.In order to verify the correctness of the theory,a calculation program of rescheduling time point is designed and verified by a concrete example.Thirdly,a genetic algorithm based on double-layer chromosome coding is proposed to solve the generation problem of flexible job shop scheduling scheme and rescheduling scheme,and the generation of scheduling scheme is designed from two aspects of algorithm and program.Finally,on the basis of the above theoretical research,a rescheduling mode selection method based on machine learning is proposed in view of the poor efficiency and low accuracy of traditional rescheduling judgment based on personnel experience.Through the comparison experiment of various machine learning modes,the machine learning method that is most suitable for the prediction and classification in this paper is obtained and improved.