A population-based, parent centric procedure for constrained real-parameter optimization

Abstract

Despite the existence of a number of procedures for constrained real-parameter optimization using evolutionary algorithms, there is still the need for a systematic and unbiased comparison of different approaches on a carefully chosen set of test problems. In this paper, we suggest a parent centric procedure for constrained real-parameter optimization. The algorithm so developed is applied to a set of 24 test problems and the results are presented. The proposed procedure is able to find the exact optimum within the specified number of function evaluations for 22 of the 24 test problems. In the remaining two problems, the proposed algorithm shows steady progress towards the respective optima, but it was unable to solve within the specified number of evaluations. It is also noteworthy that the algorithm was able to find solutions, better than the ones specified in the original problem description (http://www.ntu.edu.sg/home/EPNSugan/) for a number of test problems.