06-09-2017, 04:21 PM
Memetic algorithms (MAs) represent one of the most recent research areas in evolutionary computation. The term MA is now widely used as a synergy of evolution or any population-based approach with independent individual learning or local improvement procedures for problem-finding. Very often, MA are also known in the literature as the Baldwinian evolutionary algorithms (EAs), EA Lamarckian, cultural algorithms, or local genetic search.
Inspired by the two Darwinian principles of natural evolution and Dawkins' notion of a meme, the term "Memetic Algorithm" (MA) was introduced by Moscato in his 1989 technical report, where he saw MA as being close to a form of hybrid population-based genetic algorithm (GA) along with an individual learning procedure capable of performing local refinements. The metaphorical parallels, on the one hand, with Darwinian evolution and, on the other hand, between memes and domain-specific heuristics (local search) are captured within memetic algorithms, thus obtaining a methodology that balances well between the generality and specificity of the problem. This two-stage nature makes them a special case of double-phase evolution.
In a more diverse context, memetic algorithms are now used under various names including Hybrid Evolutionary Algorithms, Baldwinian Evolutionary Algorithms, Lamarckian Evolutionary Algorithms, Cultural Algorithms, or Local Genetic Search. In the context of complex optimization, many instantiations of memetic algorithms have been reported in a wide range of application domains, generally converging to high quality solutions more efficiently than their conventional evolutionary counterparts.
In general, the use of memetic ideas within a computational framework is called "Memetic Computing or Memetic Computation" (MC). [2] [3] With MC, the traits of universal Darwinism are more properly captured. Viewed from this perspective, MA is a more restricted notion of MC. More specifically, MA covers an MC area, in particular dealing with areas of evolutionary algorithms that marry other deterministic refinement techniques to solve optimization problems. MC extends the notion of memes to encompass conceptual entities of procedures or representations enhanced by knowledge.