Abstract

This paper describes the meta-heuristic improved harmony search algorithm (IHSA) and analyzes the performance of IHSA in solving unconstrained function minimization problems. The most important challenge to this algorithm is setting the parameters for various optimization problems. Performance of IHSA is quite sensitive to initial settings. Using Taguchi’s experimental design to arrange IHSA parameters, the performance of IHSA was analyzed to find global optima of Rosenbrock and Wood-Colville function.

Key words: Improved harmony search algorithm, Taguchi experimental design, performance of IHSA, unconstrained function.