An Improved Adaptive Harmony Search Algorithm

Abstract

Listen

Translate article

Harmony search (HS) algorithm is a heuristic optimization algorithm which is newly developed in recent years. In this paper, according to the shortcomings of the existing harmony search algorithm, an improved adaptive harmony search algorithm (IAHS) is proposed. In the IAHS algorithm, the adaptive parameters HMCR and PAR and BW are used to adjust the global and local search, so as to improve the robustness and speed of convergence of the algorithm. IAHS algorithm is tested by five standard benchmark functions and contrasted with HS、HIS and GHS algorithm. Experimental results demonstrated that the proposed IAHS algorithm has the favorable abilities of accuracy and escaping local minimums. Introduction Harmony search (HS) algorithm is a new heuristic optimization algorithm which put forward by Geem, etc in 2001 [1]. Similar with Particle swarm optimization (PSO) algorithm, HS algorithm is based on the music improvisation process. In the process, musicians repeatedly adjusting the tone of each instrument in the band, and finally get a good harmony. HS algorithm has the advantages of simple model, strong randomicity, good ergodicity and global search ability. HS and its improved algorithm has been successfully applied to many practical optimization problems [2-5], such as environmental economic load dispatch optimization, traffic path optimization, the optimization of water distribution system and fault location in distribution networks and other issues. Since the advent of harmony search algorithm, a series of achievements are made in various fields such as optimization problem: bus lines, water network design problem, the problem of reservoir scheduling, civil engineering problems. At present, this method has been widely applied in the problem of multi dimensional function optimization, pipeline optimization design, slope stability analysis etc. Research shows that the HS algorithm in solving multidimensional function optimization problems show a genetic algorithm and simulated annealing algorithm to optimize the better. However, the standard HS exists some defects such as setting BW blindly harmony memory diversity gradually dissipating with iterations, falling into local optimum easily, low accuracy and so on[6].Therefore, in order to improve the performance of the HS, this paper propose an improved adaptive harmony search algorithm (IAHS) whose parameters are adjusted adaptively. IAHS algorithm is tested by five standard benchmark functions and contrasted with HS、HIS and GHS algorithm. Experimental results demonstrated that the proposed IAHS algorithm has the favorable abilities of accuracy and escaping local minimums. Standard harmony search algorithm Harmony search (HS) algorithm was inspired by the improvising process of composing a piece of music. In the play, each player generates a tone to constitute a harmony vector. If the harmony is better, write it down, so that the next time to produce better harmony. In the algorithm, the tones of music instrument are analogous to the decision variables (i 1, 2, , n) j xi   of the optimization problem, each harmony are analogous to the solution vector ( , , , ) 1 2 j j j j x x x xn   , aesthetic evaluation are analogous to the objective function   j f x , the musicians want to find the beautiful harmonies by defined by aesthetic evaluation, the researchers want to find the global optimal solution defined by objective function .HS algorithm contains a series of optimization International Conference on Computational Science and Engineering (ICCSE 2015) © 2015. The authors Published by Atlantis Press 54 factors, such as the harmony memory (HM), harmony memory size (HMS), harmony memory consideration rate (HMCR), pitch adjusting rate (PAR), bandwidth (BW) stores the feasible solution vector represents the probability is the probability of disturbing to the Aesthetic evaluation is issued by the decision, as the value of the objective function brief description is given of the above statement Table 1 Analog elements best condition be evaluated by