Genetic Algorithm Tuning Applied to the Open Shop Scheduling Problem

Abstract

The present paper deals with the open-shop scheduling problem using a manual tuning of a genetic algorithm's parameters. A comparison has been performed between Taillard's Benchmarks for 60 instances, 2 dispatching rules and 198 variants from the GA algorithm obtained by changing the population size, the generation's number, the crossover probability, and the mutation probability. Interesting results were obtained leading to some conclusions for the best choice of the parameters.