Study on Generalized Fractal Algorithm of Global Optimization

Abstract

To promote a fractal algorithm from being suitable for global optimization only on three dimensional spaces, this paper presents a simple and convenient method for dividing an n-dimensional hypercube. A key problem is then solved to develop the fractal algorithm in a high dimensional space so that the fractal algorithm becomes a generalized global optimization algorithm. The theoritical foundation of the algorithm is set up. Simulations show the generalized fractal algorithm is effective.