Research On Lot Streaming Hybrid Flowshop Scheduling Method Based On Migrating Birds Optimization Algorithm

Hybrid Flowshop Scheduling Problem(HFSP)has important academic value and engineering significance.In order to satisfy the model of multi-variety and small-batch with batch processing,this paper is committed to introducing Lot Streaming(LS)into HFSP.Considering the static and dynamic shop environment,based on two splitting strategies(equal and inequal consistent splitting),this paper respectively develops the static scheduling and rescheduling models and studies their problem-specific knowledge and scheduling rules.Employing the Migrating Birds Optimization(MBO)as the main technical method,this paper develops scheduling algorithms incorporating problem-specific characteristics.The validity of the proposed models and algorithms is verified through the application and analysis of a practical engineering problem.The main work of this paper is as follows:For the Equal Lot Streaming Hybrid Flowshop Static Scheduling Problem(ELS_HFSP),this paper develops a Mixed Integer Linear Programming(MILP)model with the objective of minimizing the total flow time,and an EMBO(Effective MBO)based on MBO to solve the problem.In the algorithm,based on the lot streaming and objective characteristics,the “Sublot preemption” strategy,a shortest waiting time rule and a right shift and insertion mechanism are developed to ensure the lot sequence.Considering the V population structure,a combined neighborhood search strategy and competitive mechanisms are proposed to highlight the collaborative exploration.Considering the benefit mechanism in MBO,a dynamic solution acceptance criteria is developed to ensure the population diversity.To help the algorithm jump out of local optima,a scout phase based on Glover and a local search algorithm are developed.By comparing with CPLEX and the other algorithms,the effectiveness of the proposed algorithms is demonstrated.For the Equal Lot Streaming Hybrid Flowshop Rescheduling Problem(ELS_HFRP),with the consideration of the machine breakdown and ELS_HFSP solution,this paper develops a multi-objective optimization model with minimizing the total flow time and starting time deviations of sublots simultaneously.Based on EMBO and the multi-objective characteristics,with using the multi-attribute decision-making method(TOPSIS)to evaluate solutions,the MMBO(Multi-objective MMBO)is proposed.In the algorithm,a dynamic decoding process is developed with the consideration of the machine breakdown,and an improvement strategy is developed to further reduce the starting time deviations.To take advantage of the information containing in ELS_HFSP solution,a population initialization method based on Glover is proposed.In competitive mechanisms,a population shuffling process and the fast non-dominated sorting technique are introduced to adjust the population structure,and a population reproduction process is employed to improve the population quality.In scout bee phase,a local search algorithm based on Pareto characteristic is proposed to enhance the quality of candidate solution.By comparing with CPLEX and the other algorithms,the effectiveness of the proposed algorithm is demonstrated.For the Inequal Lot Streaming Hybrid Flowshop Static Scheduling Problem(ILS_HFSP),this paper develops a MILP model with the objective of minimizing the total flow time.A two level encoding is designed with the aim of determining the lot sequence and lot splitting simultaneously,and an initialization method is proposed.Based on EMBO and two level encoding,by improving the variable neighborhood descent strategy(VND)to explore the solution space sufficiently,the VMBO(VND-based MBO)is proposed.With regard to the lot splitting strategy,the “Lot preemption” is demonstrated to be more effective than “Sublot preemption” and the SWT is modified.And a selection operation and a perturbation based on block swapping are respectively introduced in the competitive mechanism and scout bee phase.By comparing with CPLEX and the other algorithms,the results demonstrate the effectiveness of the proposed algorithm.For the Inequal Lot Streaming Hybrid Flowshop Rescheduling Problem(ILS_HFRP),with the consideration of the machine breakdown and ILS_HFSP solution,this paper develops a multi-objective optimization model with minimizing the total flow time,starting time deviations of sublots and average sublot size adjustment.Based on the multi-objective characteristic and the VND strategy,with using the decomposition strategy,the MMBO/D(MMBO Based on Decomposition)is proposed.Based on the ILS_HFSP encoding and the improved strategy in ELS_HFRP,considering the lot splitting,the decoding process and the improved strategy are respectively developed.Based on the decomposition strategy,a novel benefit mechanism is developed to ensure the efficiency of the sharing process.And a variable weight strategy and a global replacement strategy are developed.Meanwhile,in competitive mechanisms,a collaborative search based on the neighboring subproblems is employed to improve the population.In scout bee phase,a novel criteria is introduced based on the varaible weight strategy to determine whether a solution is trapped into a local optimum.By comparing with the other multi-objecitve algorithms,the results demonstrate the effectiveess of the proposed algorithm.Considering a real-world engineering case in a machining workshop of an automobile production company in China,the theories and algorithms proposed in this paper are applied to the actual production scheduling.According to the actual condition in batch transferring,we respectively model the problem as ELS_HFSP and ILS_HFSP and use the CPLEX and the proposed algorithms in this paper to solve them.The results demonstrate the effectiveness the MILP models and the algorithms.Considering the machine breakdown,we respectively model the problem as ELS_HFRP and ILS_HFRP and introduce the proposed algorithms in this paper to solve them.The results demonstrate the effectiveness of the multi-objective models and the algorithms.Finally,the above work and novel points are summarized,and the future research directions are discussed.