A heuristic genetic algorithm for flowshop scheduling

Abstract

Flowshop scheduling deals with determining the optimum sequence of jobs to be processed on several machines so as to satisfy some scheduling criterion. It is NP-complete. Heuristic algorithms use problem-specific information to yield a good working solution. Genetic algorithms are stochastic, adaptive, general-purpose search heuristics based on concepts of natural evolution. We have developed a new heuristic genetic algorithm (NGA) which combines the good features of both the GA and heuristic search. The NGA is run on several problems and its performance is compared with that of the conventional genetic algorithm and the well-known NEH heuristic. The NGA is seen to perform better in almost all instances.