A Transformation Function for Finding Better Minima on Discrete Global Minimization Problems

This paper presents a new transformation function, namely T-F function, for finding better minimizers in discrete global minimization. We prove that under some general assumptions the T-F function posses the properties of both tunnelling functions and filled functions. Moreover, the new T-F function contains only one parameter and can be adjusted easily in the realization of the proposed algorithm. Numerical tests from some test functions show that our T-F function is very effective in finding better minima.