A Harmony Search with Adaptive Pitch Adjustment for Continuous Optimization

The Harmony Search is a relatively new meta-heuristic algorithm for continuous optimization, in which its concept imitates the process of music improvisation. This paper proposes an improved harmony search algorithm called Harmony Search with Adaptive Pitch Adjustment (HSAPA). The adaptive pitch adjustment scheme during a search period is inspired from the velocity clamping in particle swarm optimizations. The pitch adjustment rate is also dynamically adapted. The proposed algorithm is evaluated using thirteen wellknown benchmark functions of various characteristics. The experiment is performed using both 30 and 100 dimensions to investigate its performance at both medium and high number of decision variables. The experimental results demonstrate that HSAPA is reliably superior to all prior well-known harmony search variants and a recent widely-accepted variant of Differential Evolution algorithm. Furthermore, performance of HSAPA is shown not to be sensitive to its new parameter, which regulates the size of pitch adjustment.