Employing Genetic Algorithm and Discrete Event Simulation for Flexible Job-Shop Scheduling Problem

In this paper, a comparative study between Genetic Algorithm and Discrete Event Simulation to solve the flexible jobshop scheduling problem is presented. Two different approaches are used to generate a flexible job-shop schedule for a pharmaceutical factory X with minimum make-span which is defined as the duration required to complete all jobs. The first approach uses Genetic Algorithm to find an optimal or near-optimal solution for the flexible job-shop problem. The second approach uses Discrete Event Simulation and predefined dispatching rules to solve the flexible job-shop problem by creating a model for the pharmaceutical factory X production line. The same case study is used to evaluate the two approaches results. The Genetic Algorithm approach showed better performance compared to the Discrete Event Simulation approach for the same case study while using different dispatching rules. Both approaches showed better performance compared to basic sequential schedule.

I. Introduction

Job-Shop Scheduling Problem (JSSP) describes a scenario in which a predefined number of manufacturing jobs are scheduled on a specific number of machines. Each job consists of a set of operations which runs in a sequenced order on predefined machine types. The schedule target is to sequence all operations in a certain order so that the production line make-span is optimized. The make-span denotes the time needed to finish all scheduled jobs [1],[2]. A Flexible Job-Shop Problem (FJSP) is a deterministic extension of the JSSP where operations [3]. FJSP is solved by selecting one available machine from a certain machine-set to perform an operation with the flexibility of performing the same operation by another machine from the same set, then the basic JSSP of sequencing the jobs operations takes place.