Hybrid algorithms are powerful search algorithms obtained by the combination of metaheuristics with other optimization techniques, although the most common hybridization is to apply a local solver method within evolutionary computation algorithms. In many published works in the literature, such local solver is run in different ways, sometimes acting on the perturbed elements and other on the best ones, and this raises the question of when it is best to run the local solver and on which elements it acts best in order to improve the reliability of the algorithm. Thus, three different ways of running local search in an immune algorithm have been investigated, and well-known community detection was considered as test-problem. The three methods analyzed have been assessed with respect their effect on the performances in term of quality solution found and information gained.
Where the Local Search Affects Best in an Immune Algorithm
Scollo R. A.;Cutello V.;Pavone M.
2021-01-01
Abstract
Hybrid algorithms are powerful search algorithms obtained by the combination of metaheuristics with other optimization techniques, although the most common hybridization is to apply a local solver method within evolutionary computation algorithms. In many published works in the literature, such local solver is run in different ways, sometimes acting on the perturbed elements and other on the best ones, and this raises the question of when it is best to run the local solver and on which elements it acts best in order to improve the reliability of the algorithm. Thus, three different ways of running local search in an immune algorithm have been investigated, and well-known community detection was considered as test-problem. The three methods analyzed have been assessed with respect their effect on the performances in term of quality solution found and information gained.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.