Memetic Algorithms

Abstract

Memetic algorithms (MAs) are population‐based search strategies that have been extensively used as metaheuristics for optimization problems in a large number of domains. They are based on the synergistic combination of different algorithmic solvers, with an emphasis on hybridizations with advanced mathematical programming techniques. The synergies are obtained by balancing competitive and cooperative interactions among software agents, which are allowed to have different search strategies and sporadically interact. The paradigm was specially designed for heterogeneous distributed systems and parallel computers 20 years ago. We present a brief historical perspective on MAs, an outline of their algorithmic structural template, and some relevant design aspects of these techniques. The strong momentum of the field is further illustrated by an overview of up‐to‐date algorithmic developments of the paradigm and by highlighting some of the most recent applications published in the literature.

Iterated Local Search: Framework and Applications

“…Among the first class, which we call neighborhood-based metaheuristics, are methods like simulated annealing (SA) [16,52], tabu search (TS) [36] or guided local search (GLS) [89]. The second class comprises metaheuristics like GRASP [30], ant colony optimization (ACO) [28], evolutionary and memetic algorithms [3,67,69], scatter search [35], variable neighborhood search (VNS) [38,68] and ILS. Some metaheuristics of this second class, like evolutionary algorithms and ant colony optimization, do not necessarily make use of local search algorithms; however a local search can be embedded in them, in which case the performance is usually enhanced [28,69,70].…”

“…The second class comprises metaheuristics like GRASP [30], ant colony optimization (ACO) [28], evolutionary and memetic algorithms [3,67,69], scatter search [35], variable neighborhood search (VNS) [38,68] and ILS. Some metaheuristics of this second class, like evolutionary algorithms and ant colony optimization, do not necessarily make use of local search algorithms; however a local search can be embedded in them, in which case the performance is usually enhanced [28,69,70]. The other metaheuristics in this class explicitly use embedded local search algorithms as an essential part of their structure.…”

Heuristics for Solving Problem of Evacuating Non-Ambulatory People in a Short-Notice Disaster

“…Modern hybrid GAs further incorporate powerful local search operators as a form of mutation in order to address the situation where the population improves on average, but fails to generate near-optimal solutions. Rather than introduce small random perturbations into the offspring solution, a local search is applied to improve the solution until a local optimum is reached (Land, 1998;Moscato, 2002).…”

Memetic Algorithm with Filtering Scheme for the Minimum Weighted Edge Dominating Set Problem

“…One direction of such hybridization is to use local search which can accelerate the search process in a pure GA. This modification yields another search approach which is called the Memetic Algorithm (MA) [17].…”