A Quasi-filled Function Approach for Nonlinear Global Integer Optimization

In this paper, we propose a new quasi-filled function approach to solve box constrained discrete global optimization problem. This approach enables the objective function to move from one local optimizer to a better one at each iteration until a global optimizer or an approximate global optimizer is obtained. The properties of the proposed quasi-filled function are investigated and a solution algorithm is presented. Computational results on several testing problems are also reported.