Hybrid artificial electric field algorithm for assembly line balancing problem with equipment model selection possibility

Abstract

Existence of different models of equipment of the tasks of an assembly line has two major direct effects on the task operational times, and the equipment purchasing cost. This issue is reflected for the first time in this study in a straight assembly line balancing problem with single model product. The problem is formulated first as a non-linear model and then is linearized. Because of the NP-hard nature of the problem some meta-heuristic algorithms are proposed to solve it efficiently. For this aim a complete encoding–decoding​ scheme and the artificial electric field algorithm as a most recent algorithm of the literature and also the simulated annealing algorithm are considered. In order to fit to the problem of this study, some modified and hybrid versions of these algorithms are proposed and are tuned by the Taguchi experimental design method. A very deep computational experiments using 24 modified test problems of the literature show the superiority of the artificial electric field algorithm when is hybridized by the simulated annealing algorithm and a local search phase.