A New Approach to Evolutionary Computation: Segregative Genetic Algorithms (SEGA)

• Authors: Michael Affenzeller
• Venue: IWANN '01 Proceedings of the 6th International Work-Conference on Artificial and Natural Neural Networks: Connectionist Models of Neurons, Learning Processes and Artificial Intelligence-Part I
• Year: 2001

This paper looks upon the standard genetic algorithm as an artificial self-organizing process. With the pupose to provide concepts that make the algorithm more open for scalability on the one hand, and that fight premature convergence on the other hand, this paper presents two extensions of the standard genetic algorithm without introducing any problem specific knowledge, as done in many problem specific heuristics on the basis of genetic algorithms. In contrast to con tributions in the field of genetic algorithms that introduce new coding standards and operators for certain problems, the introduced approach should be considered as a heuristic appliable to multiple problems of combinatorial optimization, using exactly the same coding standards and operators for crossover and mutation, as done when treating a certain problem with a standard genetic algorithm. The additional aspects introduced within the scope of segregativ egenetic algorithms (SEGA) are inspired from optimization as well as from the views of bionics. In the presen paper the new algorithm is discussed for the travelling salesman problem (TSP) as a well documented instance of a multimodal combinatorial optimization prolem. In cotrast to all other evolutionary heuristics that do not use any additional problem specific knowledge, we obtaom solutions close to the best know solution for all considered benchmark problems (symmetric as well as asymmetric benchmarks) which represents a new attainment when applying evolutionary computation to the TSP.