The remarkable flexibility of evolutionary computation (EC) in handling a wide range of problems, encompassing search, optimization, and machine learning, opens up a path to attaining artificial general intelligence. However, it is clear that excessive reliance on purely stochastic evolutionary processes, with no expert guidance or external knowledge incorporation, will often lead to performance characteristics that are simply too slow for practical applications demanding near real-time operations. What is more, the randomness associated with classical evolutionary algorithms (EAs) implies that they may not be the ideal tool of choice for various applications relying on high precision and crisp performance guarantees. These observations provided the impetus for conceptualizing the memetic computation (MC) paradigm, wherein the basic mechanisms of evolution are augmented with domain-knowledge expressed as computationally encoded memes. In this chapter, we introduce what is perhaps the most recognizable algorithmic realization of MC, namely, the canonical memetic algorithm (CMA).
CITATION STYLE
Gupta, A., & Ong, Y. S. (2019). Canonical Memetic Algorithms. In Adaptation, Learning, and Optimization (Vol. 21, pp. 17–26). Springer Verlag. https://doi.org/10.1007/978-3-030-02729-2_2
Mendeley helps you to discover research relevant for your work.