206 Solving Nonlinear Programming Problems via Advanced Simulated Annealing

Deterministic global optimization methods generally have a complex optimization process, and usually make some assumptions regarding the problem to be solved. These approaches can be computationally tedious and difficult for general practitioners to implement. To overcome these drawbacks, stochastic optimization methods have been developed, such as genetic algorithms, artificial immune systems and simulated annealing (SA) algorithm. However, this study aims at SA algorithm. This study presents an advanced SA (ASA) algorithm based on penalty function. The proposed ASA algorithm is used to three benchmark nonlinear programming (NLP) problems. Numerical results indicate that the ASA algorithm can find a global solution for a NLP problem.