A hybrid honey-bees mating optimization algorithm for assembly sequence planning problem

Abstract

Assembly sequence planning (ASP) refers to taking the related constraint factors such as assembly features, assembly tools and machines into consideration to generate a low-cost feasible sequence. In this paper, a mathematical model of assembly sequence planning problem based on connectors is constructed, and a hybrid honey-bees mating optimization (HBMO) algorithm is proposed for solving this ASP problem. The proposed algorithm has two main innovative features compared to the conventional HBMO algorithm. Firstly, a crossover operator, called Multipoint Precedence Crossover (MPX), is proposed, which can avoid the generation of infeasible solutions and preserve the meaningful characteristics of the queen and broods. Secondly, worker bees utilize the simulated annealing (SA) algorithm as a local search method to improve the broods, which makes the proposed algorithm achieve the right balance between intensification and diversification. The hybrid HBMO algorithm is tested on three practical instances and compared with other approaches, such as Guided-GAs, MAs (memetic algorithm) and AIS (artificial immune systems). The superior results on these practical instances validate the effectiveness of the proposed algorithm.