An Improved Gene Expression Programming(GEP) Algorithm Based on Classification

This paper presents an improved Gene Expression Programming (GEP) algorithm, which combines the thought of classification and the original GEP operations. Meanwhile it designs a Heuristic Accelerating Searching Strategy and a Diversity Operator, which imports the thought of greedy algorithm and simulated annealing respectively. Experimental results based on comparisons between the improved GEP algorithm and the original GEP algorithm indicate that the improved GEP algorithm solves the contradiction between population diversity and algorithm convergence which has a faster convergence and better ability of searching optimization.