Harmony Search as a Metaheuristic Algorithm - Music-Inspired Harmony Search Algorithm: Theory and Applications - page 13

algorithms; and formulating completely new metaheuristic algorithms. In using the
existing algorithms, we have to identify the right algorithm for the right problem. Of-
ten, we have to change and reformulate the problem slightly and/or improve the algo-
rithm slightly in order to find the solutions more efficiently. Sometimes, we have to
develop a new algorithm from scratch to solve a tough optimization problem.
There are many ways to develop new algorithms. From the metaheuristic view-
point, the most heuristic way is probably to develop a new algorithm by hybridization.
That is to say, a new algorithm can be made based on the right combination of exist-
ing metaheuristic algorithms. For example, by combining a trajectory-type simulated
annealing with multiple agents, the parallel simulated annealing can be developed. In
the context of the HS algorithm, by the combination of HS with PSO, the global-best
harmony search has been developed [22].
As in the case of any efficient metaheuristic algorithms, the most difficult thing is
probably to find the right or optimal balance between diversity and intensity in
searching the solutions. Here, the most challenging task in developing new hybrid al-
gorithms is probably to find out the right combination of which feature/components of
existing algorithms.
A further extension of the HS algorithm will be to solve multiobjective problems
more naturally and more efficiently. At the moment, most of the existing studies,
though very complex and tough per se , have mainly focused on the optimization with
a single objective with or without a few criteria. The next challenges would be to use
the HS algorithm to solve tough multiobjective and multicriteria NP-hard optimiza-
tion problems.
Whatever the challenges will be, more HS algorithms will be applied to various prob-
lems and more systematic studies will be performed for the analysis of HS mechanism.
Also, more hybrid algorithms based on HS will be developed in the future.
References
1. Geem, Z.W., Kim, J.H., Loganathan, G.V.: A new heuristic optimization algorithm: Har-
mony search. Simulation 76, 60-68 (2001)
2. Lee, K.S., Geem, Z.W.: A new meta-heuristic algorithm for continuous engineering opti-
mization: harmony search theory and practice. Comput. Methods Appl. Mech. Engrg. 194,
3902-3933 (2005)
3. Harmony Search Algorithm (2007) (accessed December 7, 2008),
http://www.hydroteq.com
4. Yang, X.S.: Nature-inspired metaheuristic algorithms. Luniver Press (2008)
5. Glover, F., Laguna, M.: Tabu search. Kluwer Academic Publishers, Dordrecht (1997)
6. Blum, C., Roli, A.: Metaheuristics in combinatorial optimization: Overview and concep-
tual comparison. ACM Comput. Surv. 35, 268-308 (2003)
7. De Jong, K.: Evolutionary computation: a unified approach. MIT Press, Cambridge (2006)
8. Holland, J.H.: Adaptation in natural and artificial systems. The University of Michigan
Press, Ann Arbor (1975)
9. Goldberg, D.E.: Genetic algorithms in search, optimization, and machine learning. Addi-
son Wesley, Reading (1989)
10. Kirkpatrick, S., Gelatt, C.D., Vecchi, M.P.: Optimization by simulated annealing. Sci-
ence 220, 671-680 (1983)