Diversity Assessment of Multi-Objective Evolutionary Algorithms: Performance Metric and Benchmark Problems

Diversity preservation plays an important role in \nthe design of multi-objective evolutionary algorithms, but the \ndiversity performance assessment of these algorithms remains \nchallenging. To address this issue, this paper proposes a performance \nmetric and a multi-objective test suite for the diversity \nassessment of multi-objective evolutionary algorithms. The \nproposed metric assesses both the evenness and spread of a \nsolution set by projecting it to a lower-dimensional hypercube \nand calculating the “volume” of the projected solution set. The \nproposed test suite contains eight benchmark problems, which \npose stiff challenges for existing algorithms to obtain a diverse \nsolution set. Experimental studies demonstrate that the proposed \nmetric can assess the diversity of a solution set more precisely \nthan existing ones, and the proposed test suite can be used to \neffectively distinguish between algorithms with respect to their \ndiversity performance.