A gene expression programming algorithm for multiobjective site-search problem

Multiobjective site selection is a class complicated spatial analysis problem which can hardly be solved with traditional methods of Geographical Information System (GIS). In this paper we described an approach based on the gene expression programming (GEP) algorithm, with which the multiobjective site-search problems can be resolved. The validity of this method is verified by using MOP2 function, Bohachevsky function and Shubert function. By the comparison with genetic algorithms, it is concluded that the proposed GEP method using the expression trees/simple strings coding strategy can generate more approximate Pareto-front than the GAs using the linear coding method. This proposed model is finally applied to facilities optimal location search in Guangzhou.