Novel Diversity-Preservative Strategies for Genetic Algorithms and Its Application for Large-Scale Optimization

In order to increase performance of genetic algorithms, many approaches with aim of preserving diversity have been published. However, most of these approaches can be only applied to continuous optimization problems. This does not mean that genetic algorithms do not need population diversity, when they are applied to combinatorial optimization problems. In fact, defining the concept of similarity between solutions of combinatorial optimization problems, due to their apparent differences, is not straightforward. For example, for travelling salesman problem, how to measure similarity between solutions? This paper presents diversity preservative strategies which are based on similarity between solutions. These strategies not only can be applied to continuous optimization problems, but also by proposing novel semantic-oriented approaches to compute similarity between solutions of combinatorial optimization problems, it is possible to apply to combinatorial optimization problems, successfully.