A GENETIC ALGORITHM-BASED MODEL FOR JOB-SHOP FLOW TIME AND DUE-DATE SETTING IN AN ENGINEERING WORKSHOP

Abstract

Job Shop is a workshop consisting of different machines used to execute various kinds of jobs. In production generally, time is of essence. Time optimization and accurate prediction of delivery date are important to satisfying customers in a job shop. Clients desire to have their jobs concluded in good time, hence the need for schedule optimization. In this work, Genetic Algorithm is hybridized with Constructive Greedy (CG) Algorithm to achieve time optimization. Initial population is generated randomly; roulette wheel selection technique is used to rank and select the chromosomes to be mated; partially map cross-over (PMX) operator is used to combine chromosomes (solutions of the job) and CG Algorithm is adapted to carry out mutation. The research is implemented using Java programming language, Apache Tomcat 8 and web browser; and MYSQL database. This Job Shop System focused on production of bolts and nuts and the result was compared with dispatching rules such as First Come First Serve (FCFS), Shortest Processing Time (SPT) and longest Processing Time (LPT) using their makespan and mean flow time as the metrics. The result of the experiment showed that hybridizing Genetic Algorithm with Constructive Greedy algorithm produces a better result than the dispatching rules.