Optimization of a Two Sided Assembly Line using Genetic Algorithm & Discrete Event Simulation

Abstract

Two sided assembly line balancing is a complex np-hard combinatorial problem. A number of studies have been carried on one sided assembly lines, the literature is scarce on two sided assembly lines and especially type-2 problem where cycle time is to be minimized. This includes a number of constraints while allocating task to a station. Apart from the general constraints like precedence, maximum time availability and selection of a side for assembly line, the practical challenges being faced in real life scenarios are tasks positional to a station, synchronism constraints, zoning of tasks and directional zoning of tasks. A limited number of research publication on two sided assembly lines with multiple constraints have applied techniques like genetic algorithm, simulated annealing, teaching learning based optimization and ant colony algorithm. In this study proposed is a genetic algorithm with permutation based encoding applied to minimize cycle time. To mimic the real life scenario the case considered in this study is related to a motorcycle assembly line. Discrete event simulation of assembly line has been carried to compare the results before and after optimization.