Hybridmetaheuristicswithevolutionaryalgorithmsspecializinginintensificationand diversification: Overviewandprogressreport

Nowadays, apromisingwaytoobtainhybridmetaheuristicsconcernsthecombinationofseveralsearch algorithms withstrongspecializationinintensificationand/ordiversification.Theflexiblearchitectureof evolutionary algorithmsallowsspecializedmodelstobeobtainedwiththeaimofprovidingintensification and/or diversification.Theoutstandingrolethatisplayedbyevolutionaryalgorithmsatpresentjustifies the choiceoftheirspecialistapproachesassuitableingredientstobuildhybridmetaheuristics. This paperfocusesonhybridmetaheuristicswithevolutionaryalgorithmsspecializinginintensification and diversification.Wefirstgiveanoverviewoftheexistingresearchonthistopic,describingseveral instances groupedintothreecategoriesthatwereidentifiedafterreviewingspecializedliterature.Then, with theaimofcomplementingtheoverviewandprovidingadditionalresultsandinsightsonthislineof research, wepresentaninstancethatconsistsofaniteratedlocalsearchalgorithmwithanevolutionary perturbation technique.Thebenefitsoftheproposalincomparisontootheriteratedlocalsearchalgo- rithms proposedintheliteraturetodealwithbinaryoptimizationproblemsareexperimentallyshown. The goodperformanceofthereviewedapproachesandthesuitableresultsshownbyourinstanceallow an importantconclusiontobeachieved:theuseofevolutionaryalgorithmsspecializinginintensification and diversificationforbuildinghybridmetaheuristicsbecomesaprospectivelineofresearchforobtaining effective searchalgorithms.