Evolutionary Biology Evolutionary Computation
Wolfgang Banzhaf, Ting Hu
  • LAST REVIEWED: 13 December 2021
  • LAST MODIFIED: 26 November 2019
  • DOI: 10.1093/obo/9780199941728-0122


Evolutionary computation (EC) is the area of computer science and engineering that concerns itself with algorithms derived from formalizing natural evolution. This is part of a larger effort to draw inspiration from biological systems for computational purposes. Evolutionary computation methods have been used to solve optimization problems, to model systems, and to recognize patterns among other application tasks. Due to their reliance on stochasticity, they are characterized as heuristic search methods. The main features of evolutionary computation methods are their reliance on populations of searchers, the stochasticity of the search processes through mutation and recombination operations, and the application of relative strength as their selection criterion. The principle of cumulative selection allows searchers to continuously improve solutions until predefined termination criteria for the algorithms are fulfilled. The literature on evolutionary computation is comprised of a large body of proposals for algorithmic variants including hybridization schemes with other algorithms; of theoretical examinations of convergence features and other characteristics of particular variants; and of empirical studies of their performance under various testing environments, which are either constructed artificially or taken from practical applications to benchmark these variants. Furthermore, individual practical applications are published as stand-alone contributions to various fields of engineering, science, and other disciplines. Besides explicit fitness, the selection criteria for solution quality driven by external purposes like particular applications, other algorithms are studied under intrinsic selection criteria like reproductive success in an environment. Algorithms of this type come under the heading of digital or computational evolution and intend to more closely model the natural systems EC algorithms draw inspiration from. This entails studies of robustness and evolvability under various systems settings, as well as examinations of the power of algorithms to provide creative novel solutions under more-natural conditions like in an ecosystem.

General Overview

We start with a general discussion of the history of the field and recent developments. We then introduce major sources of material (textbooks, journals, conferences, websites, and software). The origin of ideas for evolutionary computation can be found in biological evolution and is derived by a method called bioinspiration, the next section. We then present key aspects of all evolutionary computation approaches, selection, representation, and variation operators. The section on coevolution explains how it allows more dynamical types of fitness definition; it is followed by a discussion of more-recent research topics. Two important concepts in evolutionary computing (robustness and evolvability of solutions) are the subject of the next section, while the more practical issue of their natural parallel nature is reviewed in the parallel implementation section. We conclude by pointing out connections with the field of artificial life and include a section on the diversity of applications of evolutionary computation methods.

back to top

Users without a subscription are not able to see the full content on this page. Please subscribe or login.

How to Subscribe

Oxford Bibliographies Online is available by subscription and perpetual access to institutions. For more information or to contact an Oxford Sales Representative click here.