This paper presents a hybrid evolutionary search method based on clusters (
HESC). The method is specifically designed to enhance the search efficiency
while alleviating the problem of premature convergence inherent in standar
d evolutionary search methods (SES). It involves the simultaneous evolution
of a main species and an additional fast mutating species, A hybrid search
method which includes a local parallel single agent search and a global mu
ltiagent evolutionary search is carried out for the main species. Effective
utilization of the search history is achieved with the clustering and trai
ning of a fuzzy ART neural network (ART NN) during the search. The advantag
es of HESC include 1) guaranteed population diversity at each generation, 2
) effective integration of local search for the exploitation of important r
egions and the global search for the exploration of the entire space, and 3
) fast exploration ability of the fast mutating species and migration from
the additional species to the main species. Those advantages have been conf
irmed with experiments in which hard optimization problems were successfull
y solved with HESC.