| Summary: | A well-known approach for solving constrained optimization problems is based on penalty functions. A penalty technique transforms the constrained problem into an unconstrained problem by penalizing the objective function when constraints are violated and then minimizing the penalty function using methods for unconstrained problems. In this paper, we analyze the implementation of a self-adaptive penalty approach, within the electromagnetism-like population-based algorithm, in which the constraints that are more difficult to be satisfied will have relatively higher penalty values. The penalties depend upon the level of constraint violation scaled by the average of the objective function values. Numerical results obtained with a collection of well-known global optimization problems are presented and a comparison with other stochastic methods is also reported.
|
|---|