Find in Library

Now, there are a lot of multiobjective evolutionary algorithms (MOEAs) available and these MOEAs argue that they are competitive. In fact, these results are generally unfair and unfaithful. In order to make fair comparison, comprehensive performance index system is established. The weights among the performance index system are solved by an adaptive differential evolution (ADE) algorithm. An approach is proposed to estimate MOEAs based on box plots and fuzzy TOPSIS. Box plots are employed to depict features of performance indicators and fuzzy TOPSIS is used to make evaluation. Experiments have been tested on IEEE CEC2009. The experiment results have revealed that the evaluation approach is effective, fair, and faithful when evaluating MOEAs.