Differential Evolution Algorithm Based on Adaptive Multi-variation Niche
The standard differential evolution algorithm(DE)has the drawbacks of weak local search ability and high parameter sensitivity.In order to improve the performance of the algorithm,an adaptive multi mutation based differential evolution algorithm(AMVDE)is designed.This algorithm first uses reverse learning strategy to improve the quality of individuals during the population initialization stage,and then increases the exploratory and robust properties of the algorithm through multiple mutation strategies and adaptive control parameter scaling.Finally,the AMVDE algorithm was applied to solve the super complex protein structure prediction problem,verifying the effectiveness and reliability of the algorithm.
differential evolution algorithmreverse learningparameter adaptationprotein structure predictioncoarse-grained energy model
万方数据
维普
