Batch Scheduling Study Of Non-equivalent Parallel Machines Based On Hybrid Discrete Particle Swarm Algorithm

Facing the era of market changes and gradual personalization of demand,the production method of multiple varieties and small lots is gradually becoming the mainstream production method.For multi-variety,small-lot production mode,batch scheduling is based on the principle of batch size,and the workpieces to be processed are divided into multiple sub-batches under the premise of meeting production requirements in order to improve equipment utilization,reduce production costs,improve production efficiency,reduce production costs and shorten production cycle time.The number of parts to be processed,the batching pattern and the processing sequence will increase the complexity of production management.Therefore,it is important to study the batch scheduling problem and find a suitable batch scheduling solution to improve equipment utilization,shorten production cycle time and reduce production cost.In this paper,based on the summary and analysis of the current status of parallel machine batch scheduling research at home and abroad,a mathematical model of multiobjective optimization is established from the actual problem of large differences in completion time and high manufacturing cost of non-equivalent parallel machines in enterprises,with the combined optimization objectives of minimizing the maximum completion time and equipment completion time balance rate.The non-equivalent parallel machine batch scheduling problem can be subdivided into two levels,one is to determine the optimal batching scheme for the workpieces to be processed,and the other is to determine the optimal production order of sub-batches on each equipment.In order to achieve global optimization so that the obtained optimal solution is closer to the true optimal solution,a centralized optimization strategy is used to optimize both the batching strategy and the subbatch ordering problem,and the same objective function is used at all levels of the algorithm as the criterion to discriminate the goodness of the feasible solution.In encoding,multidimensional encoding is used to better achieve the centralization of information,and the mapping principle is used to decode the particles.The hybrid discrete particle swarm algorithm inherits the strong global search ability of the discrete particle swarm algorithm,but also combines the local search ability of the simulated annealing algorithm to perform local search for the optimal solution,so that it has the ability to jump out of the local optimal solution.Finally,the feasibility of the hybrid discrete particle swarm algorithm is verified by comparing it with the genetic evolutionary difference algorithm and the gray wolf difference evolutionary algorithm,and the effectiveness and stability of the hybrid discrete particle swarm algorithm designed in this paper in solving the batch scheduling problem of non-equivalent parallel machines is demonstrated.And the optimization is carried out for the SMT placement line batch scheduling problem example of Changchun Continental Automotive Electronics(Changchun)Co.The research results have effective engineering applications for actual shop floor scheduling.