Research On Swarm Intelligence Algorithms For Lot-streaming Shop Scheduling Problems

With the increasing market competition and the individuation and diversification of customers'demands,the multiple-type and small-batch production mode has gradually become a mainstream mode of production.Lot-streaming shop scheduling problems exist widely in the enterprises with this mode.Therefore,the research on solving methods for these problems can be of academic and practical significance.Based on the two typical production enviroments of flowshop and jobshop,this paper studies several related lot-streaming scheduling problems by using swarm intelligence algorithms.The main works are as follows:(1)For the integrated lot-streaming flowshop scheduling problem,an improved Migrating Birds Optimization(MBO)algorithm,named IMMBO,is presented to minimize the maximum completion time(C_max).In the IMMBO,considering the characteristic of lot splitting,a harmony search based neighborhood structure is proposed,which makes good use of optimization information from the population and tunes the search scope adaptively.Besides,in view of the ease of premature convergence of the MBO algorithm,a leaping mechanism with probability is introduced to avoid being trapped in the local optimum.At last,simulation experiments show the effectiveness of the proposed algorithm.(2)For the lot-streaming distributed flowshop scheduling problem with the order constraint,a Mixed Integer Linear Programming(MILP)model with the objective of minimizing the C_max is developed.Then,three heuristics and a discrete Artificial Bee Colony(ABC)algorithm,named ORABC,are presented to solve the problem.In the ORABC,the best heuristic proposed is applied to create the initial swarm with high quality.To improve the performance of neighborhood search,a multi-neighbor strategy and a two-level search scheme are designed.Besides,considering the characteristic of the problem,a customer order based destruction-reconstruction method is put forward to generate new members.Finally,simulation experiments indicate the effectiveness of the proposed algorithm.(3)For the lot-streaming distributed heterogeneous flowshop scheduling problem with the carryover sequence-dependent setup time,an MILP model with the objective of minimizing the C_max is established.Afterwards,five heuristics and a discrete ABC(NEABC)algorithm are proposed to solve the problem.In the NEABC,a novel initialization method based on the solution created by the best heuristic presented is designed.Given the lack of interaction among individuals in the basic ABC algorithm,an effective collaboration mechanism is introduced in the onlooker bee stage.Moreover,a swarm restart strategy is employed in the scout bee stage with consideration of the special onlooker bee stage of the algorithm.Finally,simulation experiments demonstrate the effectiveness of the proposed algorithm.(4)For the lot-streaming flexible jobshop scheduling problem,a hybrid ABC(hy ABC)algorithm is presented to minimize the total flowtime.In the hy ABC,a dynamic scheme is introduced to fine-tune the search scope adaptively and thus improve search quality.In the light of poor exploitation ability of the ABC algorithm,a modified MBO(MMBO)algorithm is developed and integrated into the search process.In the MMBO,an improved downward-sharing scheme is adopted to enhance diversification of the population.Furthermore,a forward-sharing strategy is designed to take advantage of valuable information from the population.At last,simulation experiments verify the proposed algorithm is effective.(5)The theories and algorithms in this paper are applied to the real shop scheduling problems from three practical cases,and the results validate their guidance to production practices.