Integration of constraint solving techniques in genetic algorithms

A new method for constraint handling in genetic algorithms is presented in this paper. It provides a way of representing and handling constraints that is both general and problem-independent. The main idea behind this approach lies in the augmentation of genetic search with constraint solving techniques. Thus providing a general means for the handling of any kind of constraints and the incorporation of problem-specific knowledge in genetic algorithms. The application of the method is illustrated by a job shop scheduling problem. The results indicate that the method is competitive to previous problem-specific GA-based approaches when applied to scheduling problems