Abstract
Differential Evolution (DE) is a powerful stochastic algorithm for real-coded optimization. However, DE has a problem as well as other traditional stochastic optimization algorithms: that it is difficult to optimize areas that are little globally convex. Thus, it is difficult for DE and traditional algorithms to optimize some practical problem where globally convex cannot be supposed. To solve this problem, we propose Differential Evolution on Scattered Parents (DE-SP) that re-selects the individuals on each dimension when the mutant individual is calculated and some children individuals' candidates unconditionally become the children individuals. We have implemented three types of optimization experiment to verify the effectivity of DE-SP: Noisy Function 1 (NF1), that is a benchmark problem with little globally convex, Noisy Function 2 (NF2), that is the one with globally convex, and an optimization problem for bipedal robot to stand stably. Thereby, we confirmed that DE-SP was the most stable algorithm to optimize areas that is little globally convex among the comparative existing algorithms: DE, DE/nrand/1, DE/isolated/1, Hybridizing Particle Swarm Optimization with Differential Evolution, and Wavelet-Mutation-Wavelet-Crossover-Based Differential Evolution, and found the best objective function value at the practical problem.
