Intelligent Genetic Algorithm Crossover Operators for Market-Driven Design

Heuristic algorithms have been adopted as a means of developing solutions for complex problems within the design community. Previous research has looked into the implications of genetic algorithm tuning when applied to solving product line optimization problems. This study investigates the effects of developing informed heuristic operators for product line optimization problems, specifically in regards to optimizing the market share of preference of an automobile product line. Informed crossover operators constitute operators that use problem-related information to inform their actions within the algorithm. For this study, a crossover operator that alters its actions based on the relative market share of preference for each product within product lines was found to be most effective. The presented results indicate a significant improvement in computational efficiency and increases in market share of preference when compared to a standard scattered crossover approach. Future work in this subject will investigate the development of additional informed selection and mutation operators, as well as problem informed schema.