Hyper-heuristic approach: automatically designing adaptive mutation operators for evolutionary programming

Abstract Genetic programming (GP) automatically designs programs. Evolutionary (EP) is a real-valued global optimisation method. EP uses probability distribution as mutation operator, such Gaussian, Cauchy, or Lévy distribution. This study proposes hyper-heuristic approach that employs GP to design different operators for EP. At each generation, the algorithm can adaptively explore search space according historical information. The experimental results demonstrate with adaptive operators, designed by proposed hyper-heuristics, exhibits improved performance over other versions (both manually and designed). Many researchers in evolutionary computation advocate (which do adapt time) non-adaptive not alter time). core motive of this we outperform operators.