A Genetic Algorithm-Based Approach for Flexible Job Shop Scheduling

Abstract

Flexible job shop scheduling is a hard combinatorial optimization problem. This paper introduces a simulation-based Genetic Algorithm approach to solve flexible job shop scheduling problem. Four manufacturing scenarios have been considered to access the performance of a job shop with objective to minimize mean tardiness, mean flow time and makespan. Results show that multiple process plans performs better than single process plan for each job type and if only single process plan is made available, then process plan selected on the basis of minimum production time criterion yields better results than other criterion of randomly selected process plan and minimum number of set-ups. Moreover, embedding restart scheme into regular Genetic Algorithm results improvement in the fitness value.