Study On Job Shop Scheduling Optimization Method Considering The Impact Of Random Machine Failure

As a direct performer of manufacturing,the job shop not only carries a large number of production tasks,but also a place where a large amount of information meets.As the scale of production expands and the complexity of operation increases,the uncertainty in the shop manufacturing system increases sharply.As a common uncertainty factor,machine failure will not only cause deviations between the actual production process and the plan,but also cause material distribution.The chaos of the secondary plan has had a major impact on the production process.Therefore,how to effectively evaluate the performance of feasible scheduling schemes in machine random fault environment and to develop a robust scheduling scheme has become a frontier and hotspot in the field of production scheduling.Due to the randomness of machine faults,the complexity of production tasks,and the diversity of scheduling objectives,the complexity and difficulty of this kind of shop scheduling problem are determined.On the basis of considering the dynamics,complexity and target diversity of the shop scheduling problem,the paper makes a deep research on the scheduling performance evaluation and robust scheduling method in the machine fault environment.The main research work of this paper is:(1).By analyzing the relationship between processes,a process-related network with processes as nodes,process-constrained relationships as edges,and time-to-process time as weights is constructed to provide a new way of describing the relationships between processes.(2).For the performance evaluation method of the scheduling scheme in the machine fault environment,two types of robustness indicators are defined.Starting from the process-related network,by mapping the machine fault to the process level,analyzing the propagation effect of the single-process fault on the associated network,and then evaluating the comprehensive effect of multi-process fault propagation,two new robust measure methods are proposed.And the solution algorithm is given.Simulation experiments on 25 typical job shop cases show that the two measures are faster than Monte Carlo simulation and more accurate than the three alternative robustness measures.(3).Assuming that the machine fault obeys the Weibull distribution,a multi-objective scheduling model with performance robustness and stable robustness considering scheduling is established.By analyzing the internal structure of the scheduling scheme,machine fault information and the expected delay of the process Relationship,the accurate evaluation method of the scheduling scheme performance is given;combined with the expected delay of the process and the impact of the evaluation process on the scheduling performance of the backward associated process,the concept of process influence is proposed;by adding non-forced idle after a few influential processes Time,achieving significant stability improvements with less performance loss;and designing improved genetic algorithm solutions.Through the simulation experiments of 25 typical job shop cases under different parameters,the experimental data show that with the increase of fault repair time,the performance robustness of the scheduling scheme degrades faster than the stable robustness.By adding unforced buffering time in a few influential processes,it is possible to achieve higher stability with higher performance loss in exchange for higher performance.