A DERIVATIVE-FREE EXPLORATORY TOOL FOR FUNCTION MINIMIZATION BASED ON TABU SEARCH

Tabu search is a relatively new optimisation procedure that has the ability to escape local optima in the search for a global optimum. In general its applications have concentrated on combinatorial problems that are NP-hard. In this article an implementation of tabu search is described for minimising unconstrained functions in a non-combinatorial context. The algorithm was tested on several functions (some standard) and shown to be useful and effective, particularly as an interactive exploratory tool.