A Hybrid Evolutionary Optimization Model for Solving Job Shop Scheduling Problem using GA and SA

摘要

The heuristic optimization techniques were commonly used in solving several optimization problems. The present work aims to develop a hybrid algorithm to solve the scheduling optimization problem of JSSP. There are different variants of these algorithms that were addressed in several previous works. The impacts of these two kinds (Genetic Algorithm (GA) and Simulated Annealing (SA) based optimization model) of initial condition on the performance of these two algorithms were studied using the convergence curve and the achieved makespan. Even though genetic algorithm performed better than other evolutionary algorithms, it has some weakness. During running GA, sometimes, it will produce same result without any improvement. SA has a mechanism to overcome from that situation. During SA, if same result will be repeated, then it is rapidly changing the change in temperature variable and re-initiates another random search. By using this feature of SA, it has been implemented a hybrid based evolutionary model for solving JSSP by improving GA. Comparison has been made with the performance of the proposed SA-GA-Hybrid model with GA as well as SA.