Maumita Bhattacharya
Charles Sturt University, Australia
Download article
Published in: Nordic MPS 2004. The Ninth Meeting of the Nordic Section of the Mathematical Programming Society
Linköping Electronic Conference Proceedings 14:7, p.
Published: 2004-12-28
ISBN:
ISSN: 1650-3686 (print), 1650-3740 (online)
Evolutionary algorithms (EA) have been long accepted as efficient global optimizers. Given a search space S and an objective function g defined on it; the problem is to find the global maximum (or minimum) of g in S. To apply EA’s heuristic search; the coding function or representation ? is created; that partially maps S to the finite chromosome space C. The genetic operators are used to create new solutions such that Cn ? Cm.
However; as the evolutionary search progresses; it is important to avoid reaching a state where the genetic operators can no longer produce superior offspring; prematurely. This is likely to occur when the search space reaches a homogeneous or near-homogeneous configuration converging to a local optimal solution. Maintaining a certain degree of population diversity is widely believed to help curb this problem. This paper discusses the problem of premature convergence related to EA based optimization. A novel technique is presented; that uses informed genetic operations to reach promising; but un/under-explored areas of the search space; while discouraging local convergence; to curb premature convergence. Elitism is used at a different level aiming at convergence. The proposed technique’s improved performance in terms solution precision and convergence characteristics is observed on a number of benchmark test functions with a genetic algorithm (GA) implementation.
