Genetic Algorithms and General
Validity measures how well a string satisses the constraints and the structure of genetic algorithm is modiied to incorporate the validity of strings to show that this theory has promise.
Abstract
Genetic algorithms, while robust solution space searching methods, do not perform well when faced with general constraints. In some applications, for example the Traveling Sales-person Problem, problem-speciic operators and codings can be developed. However, general constraints are still problematic. Described here are general methods to nd an optimal or near-optimal solution to a problem while violating few if any constraints. We start by deen-ing a function which measures the validity of a string. Validity measures how well a string satisses the constraints. A method to approximate validity is also given. The structure of genetic algorithm is modiied to incorporate the validity of strings. Preliminary experiments are performed and the results show that this theory has promise.