Evaluation of the constraint method-based multiobjective evolutionary algorithm (CMEA) for a three-objective optimization problem
Sujay V. Kumar, S. Ranji Ranjithan · 2002
This paper presents a systematic comparative study of CMEA (constraint method-based multiobjective evolutionary algorithm) with several other commonly reported mulitobjec-tive evolutionary algorithms (MOEAs) in solving a three-objective optimization problem. The best estimate of the noninferior space was also obtained by solving this multiobjective (MO) problem using a binary linear programming procedure. Several quantitative metrics are used to compare the noninferior solutions with respect to relative accuracy, as well as spread and distribution of solutions in the noninferior space. Results based on multiple random trials of the MOEAs indicate that overall CMEA performs better than the other MOEAs for this three-objective problem.