Finding global minimum using filled function method

Abstract

Filled function method is an optimization method for finding global minimizers. Filled function method is a combination of a local search in findings local solutions as well as global solution. It is basically a construction and eventually the inclusion of an auxiliary function called the filled function into the algorithm. Optimizing the objective function at an initial point will only yield a local minimizer. By using the auxiliary function, the local minimizer is shifted to a new lower basin of the objective function. The shifted point is the new initial solution for the local search to find the next local minimizer, where the function value is lower. The process continued until the global minimizer is achieved. This research used several test functions to examine the effectiveness of the method in finding global solution. The results show that this method works successfully and further research directions are discussed.