Research On Flow Shop Scheduling With Lot Streaming In Workshop Under Uncertain Environment

The manufacturing industry is transforming and upgrading to the direction of intelligent manufacturing under the background of"Made in China 2025",and the production line of the manufacturing workshop is also becoming more and more intelligent and informatized.In this context,reasonable and efficient workshop production scheduling arrangements play a vital role in the efficient operation of the production line.The production scheduling problem of multiple types with small batches is one of the key problems in manufacturing at present.This paper takes the expansion of the flexible flow-shop scheduling problem,that is,the lot-streaming scheduling problem as the research direction,and considers the random arrival of the workpiece in the workshop and the setup time between the machines.The problem includes how to divide each kind of workpiece into sub-lots to be processed,and the scheduling of all sub-lots.The main work is as follows:（1）This paper takes minimizing the makespan as the optimization goal.Based on the scenario,the Sample Average Appropriation（SAA）method is used to establish the stochastic flexible flow shop scheduling with lot-streaming model MILP_SAA,which takes the setup time and the random arrival of the workpiece into account.It can be divided into several sub-lots with different batches to enter the machine for processing.To solve the problem,the batch plan of the workpiece and the sub-lot processing sequence must be given at the same time.（2）The Benders’decomposition algorithm is used to solve the model MILP_SAA accurately,and three acceleration strategies are designed to improve the efficiency of the Benders’decomposition algorithm.Numerical experiments are also designed to compare with the results directly solved by the CPLEX solver.The results verify the Benders’decomposition algorithm’s effectiveness and superiority.（3）A tabu search algorithm is designed to realize the optimal solution of the two problems of batching and scheduling in a larger scale of the problem,and the MILP_SAA was simplified,the fitness value of the solution could be quickly obtained through the Benders’decomposition algorithm.The results of the tabu search algorithm and the genetic algorithm are compared to verify the performance of the designed tabu search algorithm.The results show that the tabu search algorithm has better optimization ability and better convergence performance.（4）In order to solve the stochastic lot-streaming scheduling problem in a larger scale further,the simplified MILP_SAA was relaxed,and four rounding strategies are designed to adjust the linear solution to an integer solution that meets the requirements.The four strategies are compared through numerical experiments,and it is proved that the strategy of assigning by distance has stronger optimizing ability under normal circumstances.（5）The research scope of the problem was also extended to continuous sub-lots,and a random lot-streaming scheduling model with continuous sub-lots is established.According to the difference methods to decide the sub-problems and the master problem,two sets of Benders decomposition algorithms are designed to achieve the accurate solution of the model.The experimental results prove that dividing continuous variables into the master problem is more conducive to problem solving.