A fusion of crossover and local search

It is well known that genetic algorithms (GAs) are not well suited for fine-tuning structures that are very close to optimal solutions and that it is essential to incorporate local search methods, such as neighborhood search, into GAs. This paper explores the use of a new GA operator, called multi-step crossover fusion (MSXF), which combines a crossover operator with a neighborhood search algorithm. MSXF performs a local search essentially in the region within the search space between parent solutions to find a locally optimal solution that inherits the parents' characteristics. GA/MSXF was applied to-job-shop scheduling problem. Experiments using benchmark problems show promising GA/MSXF performance even with a small population.