A New MultiObjective Genetic Algorithm for Assembly Line Balancing

Abstract

The aim of this work is to enable a step towards a selfadapting digital toolset for manufacturing planning focusing on minimally constrained assembly line balancing. The approach includes the simultaneous definition of the optimum number of workstations, the optimum cycle time and the assignment of tasks to workstations. A bespoke genetic algorithm (GENALSAS) is proposed and demonstrated which focuses on examining the simple assembly line balancing problem (SALBP). The proposed genetic algorithm (GA) has been shown to consistently deliver detailed production plans for SALBP problem forms with minimum inputs. Neither the number of workstations nor the system cycle time is assumed/fixed as in previous work in the field. The work simultaneously attains better performing solutions compared with previous studies both in terms of time to converge and the quality of the solution.