Evolving SSE: a Stochastic Schemata Exploiter

This paper proposes Stochastic Schemata Exploiter (SSE), a new population-oriented search scheme which employs a schemata processing mechanism similar to the one used in genetic algorithm (GA). Compared with basic GA, SSE has the following two features: first, SSE exploits more the ability of local search of schemata processing while GA puts more emphasis on the ability of global search. Second, SSE reduces the number of control parameters which are in many cases problem dependent and should be determined heuristically. Because of these features, SSE is more suitable than basic GA to be combined with other advanced schemes such as niching or to be incorporated into hybrid search. In the paper, the advantage of SSE is demonstrated using GA-easy and GA-hard test functions