Real-Coded Genetic Algorithm for Solving Generalized Polynomial Programming Problems

Abstract

Generalized polynomial programming (GPP) is a nonlinear programming (NLP) method based on a nonconvex objective function, which is subject to nonconvex inequality constraints. Hence, a GPP problem has multiple local optima in its constrained solution space. General NLP techniques use local optimization, and therefore do not easily solve GPP problems. Some deterministic global optimization approaches have been developed to overcome this drawback of NLP methods. Although these approaches yield a global solution to a GPP problem, they can be mathematically tedious. Therefore, this study presents a real-coded genetic algorithm (RGA), which is a stochastic global optimization method, to find a global solution to a GPP problem. The proposed RGA is used to solve a set of GPP problems. The best solution obtained by the RGA is compared with the known global solution to each test problem. Numerical results show that the proposed RGA converges to a global solution to a GPP problem.