Resolução de problemas de otimização com restrições de igualdade e desigualdade utilizando a Inicialização Global Topográfica

Abstract

The constrained optimization methods, developed from the classical mathematical analysis, are generally very efficient numerical algorithms. However, such methods strongly depend on the location of the starting points. In this work we use the Topographical Initialization heuristic to generate adequate starting points for the local search method in order to solve global optimization problems with equality and inequality constraints. To perform the local search, we used the Feasible Direction Interior Point  Algorithm (FDIPA). We performed computational experiments using 8 problems in order to test the method performance. The results show that the present methodology is feasible and efficient for the global minimization of functions subject to restrictions of equality and inequality.