Abstract:
To address the premature convergence in the searching process of differential evolution, a chaotic migration-based multi-population differential evolution (CMMPDE) is proposed. In this algorithm, the asynchronic migration of individuals during parallel evolution is guided by a chaotic migration sequence. Because the sequence is ergodic and stochastic , information exchanging among sub-populations is ensured to be efficient and sufficient . Simulation study of CMMPDE and its application of PID control parameters optimization have proved its capability of strong global search, superiority to SDE and high immunity against premature convergence.