Research On Robust Rescheduling Of Discrete Manufacturing Enterprises Under Machine Faults

Discrete manufacturing is one of the main production modes of the manufacturing industry.Its production method is order-oriented pulling production with characteristics of multiple varieties,small batches and high flexibility.The production process is in a dynamic environment and faces many disturbing factors.The complexity of discrete manufacturing systems puts higher requirements on production scheduling management.How to quickly respond to disturbances in the production process and obtain the optimal re-scheduling plan has become a major issue for discrete manufacturing system.Scheduling is to make the tasks and machines of the manufacturing system meet the relevant constraints under the conditions of existing production resource and finally make the evaluation index reach the overall optimum.Therefore,this thesis focuses on the rescheduling problem of discrete manufacturing system.The main research content:Firstly,the overall research on the scheduling problem of discrete manufacturing system is carried out.Research status of manufacturing system scheduling problems is summarized.The current research achievements and deficiencies of the scheduling and re-scheduling problems is analyzed.The urgency of the discrete manufacturing system scheduling problem is pointed out.The scheduling characteristics of discrete manufacturing system is analyzed.The method of judgment the uncertainty of disturbance and the solution algorithm of the scheduling model are studied.Therefore,the research framework of the discrete manufacturing system scheduling problem is established.Secondly,disturbance factors of discrete manufacturing systems are analyzed and disturbance factors are divided into burst-type disturbances and cumulative-type disturbances,then the re-scheduling response strategy of discrete manufacturing systems is proposed.Rescheduled performance evaluation indicators are established.According to the constraints of discrete manufacturing systems,an initial scheduling model for discrete manufacturing systems is constructed.A two-layer coding genetic algorithm is proposed,and the initial population,fitness function and genetic operator are designed to improve the calculation speed and efficiency of the algorithm.Then,according to the uncertainty of the disturbance degree of the machine failure,the disturbance degree judgment index is designed and the probabilistic neural network input parameter matrix is established to judge the degree of disturbance.Combining with the idea of right shift re-scheduling,right-shift insertion re-scheduling is designed.With the main objective of method adaptability,complete re-scheduling is designed.Combining the disturbance degree judgment model,a hybrid re-scheduling method based on fuzzy judgment is proposed.Finally,combining with the actual production process of Y company,the probabilistic neural network and double-level coding genetic algorithm are programmed in MATLAB software,and the feasibility and effectiveness of the designed re-scheduling strategy and hybrid re-scheduling method are verified.