Multi-objective optimization of mixed model assembly line balancing in an assemble-to-order industry with stochastic environment

Abstract

The objective of this research is to propose a methodology for multi-objective optimization of a mixed-model assembly line balancing problem with the stochastic environment. To do this a mathematical model representing the problems at hand is developed with objectives of minimizing cycle time and minimization of the number of workstations (which is of Type-E ALB problem). And two optimization meta-heuristics are considered to solve it, namely, Non-Dominated Sorting Genetic Algorithm- II (NSGA-II) and Multi-Objective Genetic Algorithm (MOGA). To test the performance of the algorithms three different size standard problems in Assemble-to-order types of industry are taken and five demand arrival scenarios are considered to incorporate the stochastic nature of the demand arrival for each model in all problems. Both the algorithms are coded and run using MATLAB® 2013a and are compared based on different performance measures. The results indicated that MOGA outperformed NSGA-II in most of the test problems. Nevertheless, both algorithms have resulted in significant improvements in the performance measures in Assemble-to-order types of industry dataset compared to the existing line configuration.
 Keywords: Assembly Line, Multi-objective optimization, Single model, mixed-model, stochastic environment, Genetic Algorithm

Choosing the optimal balance for the production assembly lines at the Volvo car manufacturing plant in Babylon

“…5. Efficiency: It is the ratio between the total station time to the product of cycle time and the number of workstations, represented as a percentage, and this shows the relative efficiency of using the line (Legesse, et al, 2020, 95) [11] , and production efficiency is defined as the proper use of resources to achieve the desired goal. By reducing costs, time and efforts, ensuring access to the highest number of products in exchange for a small number of inputs (Bilal and Rayes, 2023, 124) [15] .…”