Abstract

A global optimization method for continuous design variables called as Generalized Random Tunneling Algorithm is proposed. Many global optimization methods have been proposed for the unconstrained optimization problem which is to find global minimum subject to the side constraints only. Proposed method is called as "Generalized" random tunneling algorithm because this method can treat the inequality constraints as well as the side constraints. This method consists of three phases. That is, minimization phase, tunneling phase, and constraints phase. By iterating these phases, global minimum is found. The characteristics of mathematical programming and heuristic approaches are included in the proposed method. Global minimum which may lie in the boundary of constraints is easily found by proposed method. Proposed method is applied mathematical and structural optimization problems. Through numerical examples, the effectiveness and validity of proposed method have been confirmed