Research on adjustment strategy of PAR in harmony search algorithm
Changming Xing, Lin Yang · 2012
As a meta-heuristic algorithm, Harmony Search (HS) can be used to solve lots of NP-hard problems. The pitch adjustment rate (PAR) in HS algorithm is very important parameter in balancing local exploitation and the global exploration. In this paper, three new PAR adjustment strategies are presented for enhance the global optimality of HS, which follow small PAR values with large BW values in early generations, and large PAR values with small BW values in final generations. Finally, the computational simulations and comparisons are carried out by employing six benchmark problems from related literature. The experimental results indicate that using convex function adjust the PAR could improve the performance of HS algorithm. Conversely, using the concave function adjust the PAR reduce the performance of HS algorithm.