A Filled Function for Non-smooth Box-constrained Global Optimization and Its Application

Abstract

This paper constructs a filled function for non-smooth box-constrained global minimization, investigates its properties and designs a corresponding algorithm. The constructed new filled function contains only one parameter and it has more merits than those with several parameters. Moreover, the iterative operations of the proposed filled function algorithm can be easily implemented. we also perform numerical experiments to show the efficiency of the proposed approach.