Assembly sequence planning based on hybrid artificial bee colony algorithm

Abstract

Intelligent algorithm provides a promising approach for solving the Assembly Sequence Planning (ASP) problem on complex products, but there is still challenge in finding best solutions efficiently. In this paper, the artificial bee colony algorithm is modified to deal with this challenge. The algorithm is modified from four aspects. First, for the phase that employed bee works, a simulated annealing operator is introduced to enrich the diversity of nectar sources and to enhance the local searching ability. Secondly, in order to prevent the swarm from falling into local optimal solutions quickly, a tournament selection mechanism is introduced for the onlooker bees to choose the food source. Thirdly, for the phase that scout bee works, a learning mechanism is introduced to improve the quality of new generated food sources and to increase the convergence speed of the algorithm. Finally, a fitness function based on the evaluation indexes of assemblies is proposed to evaluate and select nectar sources. The experimental results show that the modified algorithm is effective and efficient for the ASP problem.