Improving differential evolution using a best discarded vector selection strategy

2022 
In this study, a selection strategy based on discarded trial vectors was constructed to improve the performance of the differential evolution (DE) algorithm. The proposed strategy is called the best discarded vector selection (BDVS) strategy. The uniqueness of BDVS is that the best discarded trial vector for an individual and the best vector for the individual are used to improve the individual. First, an is constructed to record the best discarded trial vectors for each individual. During the survivor selection operation, if a trial vector is worse than the target vector and the target vector is in a stagnant state, then the target vector is replaced with the best vector of the target vector or a vector chosen at random from the external archive. BDVS is a general framework that can easily be integrated into various DE algorithms. In our experiments, BDVS was applied to eight DE algorithms, which were then compared to their original algorithms. BDVS was verified using the 2013 IEEE Congress on Evolutionary Computation (CEC 2013), CEC 2014, and CEC 2017 benchmark test sets and 88 test functions were considered. The experimental results demonstrate that BDVS can significantly improve DE performance.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []