RESEARCH
Optimization algorithms(Meta-heuristics)
■ Optimization Algorithms (Meta-heuristics)
Optimization is the process of selecting a bestelement from some sets of available alternatives according to some criteria. Ineach iteration of the optimization process, choosing the values from within anallowed set is done systematically until the minimum or maximum result isreached or when the stopping criterion is met. Meta-heuristic algorithms arewell-known optimization algorithms which can solve optimization problems withsatisfying a set of constraints.
■ Harmony Search Algorithm (website)
TheHarmony Search (HS) algorithm was developed by Prof. Kim and Dr. Geem (a formermember of Hydrosystems lab) is one of the most recently developed optimizationalgorithm and at the same time, it is one the most efficient algorithm in thefield of combinatorial optimization.
The HS algorithm was inspired by musical performanceprocess involving searching for a best harmony. When the orchestra membersfirst practice, there can be the good and bad harmonies. While practicing, badharmonies are removed, and finally it creates the fantastic harmony (optimal solution).Since the HS algorithm was first developed in 2001, the HS was initiallyapplied in well-known benchmark problem (e.g., De Jong functions), parameterestimation, travelling salesman problem, and optimal design of water distributionnetworks (WDNs).
● Z.W. Geem, J.H. Kim,G.V. Loganathan, "A New Heuristic Optimization Algorithm: Harmony Search."Simulation, Vol.76, No 2, pp 60 ~ 68, Feb 2001.
● J.H. Kim, Z.W. Geem,E.S. Kim, "Parameter Estimation of the Nonlinear Muskingum Model usingHarmony Search." Journal of the American Water Resources Association,Vol.37, No.5, pp 1131 ~1138, Oct 2001.
In the HS’s applications of practical problems,however, studies about water resources and structural engineering hold largemajority in the early stage because the major field of the developer is a civilengineering (CE). Since 2001, it has been applied to various engineeringproblems, and the international attention on it has rapidly increased.
In each iteration of the HS algorithm, a new solutionis generated either by random search (RS) or by harmony memory consideration(HMC) followed by pitch adjusting (PA). HMC is performed with a probability ofharmony memory considering rate (HMCR) while PA is processed with a probabilityof a pitch adjustment rate (PAR). In addition, harmony memory size (HMS) indicatingthe size of memory space (harmony memory, HM) and band width (BW) defining theallowable range of PA are used as the parameters.
In the early stage, some studies attempted to modifythe parameters called the modified HS, revised HS, and improved HS. Recently,self-adaptive or parameter setting free versions of HS have been developed inan effort to avoid providing initial values of decision variables as well aschanging the parameter values manually in the course of the search. Recently,hybrid HSs and multi-objective HSs has been developing.
● Revised HarmonySearch (ReHS) (Paik et al., 2002) [Link_paper]
● Modified HarmonySearch (MHS) (Paik et al., 2005) [Link_paper]
● Particle-SwarmHarmony Search (PSHS) (Geem, 2009) [Link_paper]
● Parameter SettingFree Harmony Search (PSF-HS) (Geem and Sim, 2010) [Link_paper]
● Smallest Small World CellularHarmony Search (Im et al., 2013) [Link_paper]
In applications aspects, the HS was initially appliedin benchmark and simple engineering problems to verify the applicability of thealgorithm itself. Recently, however, it has been used and utilized in thereal-world problems. The application results show that HS can be effectivelyused as tool for optimization problems. Therefore, it can be recommended as theworld-wide optimization techniques, and it also can be used in wide range ofengineering field.
■ Water Cycle Algorithm
Water Cycle Algorithm (WCA) as a recently developedoptimizer is introduced by Eskandar et al. (2012). The fundamental concepts andideas which underlie the WCA is inspired by nature and based on the observationof water cycle process and how rivers and streams flow to the sea in the realworld. A comparative study has been carried out to show the effectiveness ofthe WCA over other well-known optimizers in terms of computational effort(measures as number of function evaluations) and function value (accuracy) fortruss structures, unconstrained, constrained engineering design problems, and multi-objectiveoptimization problems. Below is the schematic view of the processes for theWCA, where circles, stars, and the diamond correspond to streams, rivers, andsea, respectively.
Mine Blast Algorithm (MBA) is a recently developedoptimization method used for handling complex optimization problems introducedby Sadollah et al. (2012). The fundamental concepts and ideas of the MBA arederived by the explosion of mines in real world. The efficiency of the proposedoptimizer is tested via the optimization of several truss structures havingdiscrete variables, unconstrained and constrained engineering design problems,and multi-objective optimization problems.
The performance of the MBA is compared with several well-knownmetaheuristic algorithms given in the literature. The obtained optimizationresults show that the MBA is able to provide faster convergence rate and alsomanages to achieve better optimal solutions compared to the other efficientoptimizers. Below is the schematic view of the processes (concentric lines for explorationphase and solid lines for exploitation phase) for the MBA.
■ Related ResearchProject
(Establishment of Original Technology and Globalization Strategy ofHarmony Search Algorithm)
Hydrosystems laboratoryhas researched about optimization algorithms in, ‘Establishment of OriginalTechnology and Globalization Strategy of Harmony Search Algorithm’ projectsupported by the National Research Foundation, Korea.
Thegoal of this project is to re-establish the core technology of HS algorithmthat started earning a world-wide attention and developed in the laboratory. Itis also aiming to create an algorithm that is comparable to the GeneticAlgorithm (GA) in the middle of the spotlight as a global optimizationmethodology.
→ Comparison andanalysis among meta-heuristic algorithm
→ Development of techniqueto improve HS algorithm
→ Development of benchmarkproblems in water resources engineering field
→ Globalization of HS algorithmby initiatives and strategies
→ Development multi-objectiveHS and hybrid HS
→ Development andapplication of improved algorithm to water resources engineering benchmarkproblems
For theestablishment of original technology and globalization, Hydrosystems laboratoryhosted the 1st Symposium on Harmony Search Algorithm in 2014. The 2nd InternationalConference of Harmony Search Algorithm 2015 (ICHSA 2015), held in KoreaUniversity in Seoul, South Korea from 19 to 21 August, 2015, was intended to bean international forum for researchers in the area of developing, design,variants, and hybrid methods of HS algorithm.
■ Research Achievements
● 1st Symposium onHarmony Search Algorithm
The First Symposiumon Harmony Search Algorithm was held on 13 ~ 14th of February 2014 at Anamcampus of Korea University in Seoul, South Korea. It was a good opportunity tointroduce various aspects of Harmony Search Algorithm to graduate students aswell as professors in various disciplines. In addition, it was a showcase ofthe research capabilities of teams and can bring ideas for futurecollaborations among them. A variety of presentations was done in the symposiumby 7 presenters from England, EU, Spain, Iran and Korea.
- “Harmony Search:Past, Present & Future” by Zong Woo Geem (Gachon University, Korea)
- “Analysis ofHarmony Search and Other Metaheuristic Algorithms” by Xin-She Yang (Universityof Cambridge, England)
- “The Story ofSisyphus: Harmony Search Algorithm Promoting Engineering Optimization” byIoannis Kougias (Institute for Energy and Transport, European Commission)
- “GroundwaterApplications of Harmony Search” by Mustafa T. Ayvaz (Pamukkale University,Turkey)
- “Recently DevelopedMetaheuristic Optimizers and Their Applications in Water and Civil Engineering”by Ali Sadollah (Korea University, Korea)
- “RecentApplications of Harmony Search to Urban Road Network Planning, IndoorLocalization and Telecommunications” by Javier Del Ser Lorente (TecnaliaResearch & Innovation, Spain)
- “Improving theGlobal and Local Search Capability of Harmony Search” by Joong Hoon Kim (KoreaUniversity, Korea)
Discussion on how tomanage regular meetings for the algorithm user group and field trip introducingthe Seoul city was also scheduled at the end of the symposium.
● 2nd InternationalConference of Harmony Search Algorithm 2015 (ICHSA 2015)
The Second International Conference on Harmony Search Algorithm (ICHSA2015) is held on 19th ~ 21st of August 2015 at Anam campus of Korea Universityin Seoul, South Korea. The ICHSA 2015 is intended to be an international forumfor researchers in the area of developing, design, variants, and hybrid methodsof Harmony Search Algorithm (HSA). As a result, this conference can be servedas a valuable place to create and exchange new ideas from various field. . Inaddition, it made a future research collaborations with sharing their own idea.A variety of presentations was done in the conference by 42 presenters from China,India, Iran, Malaysia, Saudi Arabia, Taiwan, Spain, Hungary, South Africa,Central Africa, and Korea and the 85 participants are enjoy this conference. Theexperts in soft computing and optimization, Hojjat Adeli (The Ohio StateUniversity), Kusum Deep (Indian Institute of Technology Roorkee), PonnuthuraiNagaratnam Suganthan (Nanyang Technological University), Ling Wang (TsinghuaUniversity), are invited to International Conference of Harmony SearchAlgorithm 2015 as keynote speakers.
- “Computational Intelligence in Engineeringand Medicine” by Hojjat Adeli
- “Can Nature-Inspired Optimization Techniquessolve complicated real-life problems ?” by Kusum Deep
- “Numerical Optimization by Nature InspiredAlgorithms” by Ponnuthurai Nagaratnam Suganthan
- “Computational Intelligence BasedScheduling” by Ling Wang
The International Steering Committee discussed on planning the next conference. We look forward to seeing again in next 3rd International Conference on Harmony Search Algorithm.