A Novel Membrane Algorithm Based on Differential Evolution for Numerical Optimization

摘要

This paper presents a novel membrane algorithm, called DEPS, for numerical optimization. DEPS is an appropriate combination of a differential evolution algorithm, a local search and P systems. In this algorithm, the hierarchical structure of cell-like P systems is used to organize the objects consisting of real-valued strings and the rules which are composed of mutation, crossover and selection operations in elementary membranes, a local search in the skin membrane and transformation/communication-like rules in P systems. The effectiveness of the algorithm is tested on extensive numerical optimization experiments. In what follows DEPS is applied to solve a real-world problem, time-frequency atom decomposition. Experimental results show that DEPS performs better than its counterpart differential evolution algorithm.