Global Optimization by Generalized Random Tunneling Algorithm (2nd Report, Examination on the Accuracy of Solution and its Efficiency)
Satoshi KITAYAMA, Koetsu YAMAZAKI · TRANSACTIONS OF THE JAPAN SOCIETY OF MECHANICAL ENGINEERS Series A · 2004
In constrained optimization problems, many heuristic methods treat constraint functions as penalty function, and it is well known that penalty parameters depend on the problems. Generalized Random Tunneling Algorithm (GRTA) for global optimization has been proposed in the first report. GRTA does not need penalty parameters, and is a simple method. In this paper the accuracy of solutions and the efficiency between GRTA and heuristic methods are discussed through numerical examples. In numerical examples, some mathematical problems and structural optimization problem are treated. In case of mathematical problems, there are little differences of the accuracy of solution and the results between GRTA and heuristic methods. However, the efficiency of GRTA is excelent to find global minimum. Moreover there is big difference of the result in case of structural optimization problem. As a result, the validity and efficiency of GRTA are confirmed.