Estimation of evolvability genetic algorithm and dynamic environments

Abstract

This article investigates the of applicability of adding evolvability promoting mechanisms to a genetic algorithm to enhance its ability to handle perpetually novel dynamic environments, especially one that has stationary periods allowing the Genetic Algorithm (GA) to converge on a temporary global optimum. We utilize both biological and evolutionary computation (EC) definitions of evolvability to create two measures: one based on the improvements in fitness; the other based on the amount of genotypic change. These two evolvability measures are used alongside fitness to modify how selection proceeds in the GA. We call this modified GA the Estimation of Evolvability Genetic Algorithm (EEGA). When tested against a regular GA (with random immigrants), the EEGA is able to track the global optimum more closely than the GA during the dynamic period. Unlike most GA extensions, the EEGA works effectively at a lower level of diversity than does the GA, showing that it is the quality of the diverse members in the population and not just the quantity that helps the GA evolve.

Bayesian Inference to Sustain Evolvability in Genetic Programming

“…Wang and Wineberg [14], suggested two measures of evolvability one based on fitness improvement and the other based on the amount of genotypic change. The authors divided the population into three sub-populations, where the size of each sub-population is determined dynamically.…”

Evolutionary Dynamic Optimization: Methodologies

“…Another approach to maintain diversity is to reward individuals that are genetically different from their parents [110]. In this approach, in addition to a regular population, the algorithm maintains an additional population where individuals are selected based on their Hamming distance to their parents (to promote diversity) and another population where individuals are selected based on their fitness improvement compared to their parents (to promote exploitation).…”

Genetic Programming — Introduction, Applications, Theory and Open Issues

“…Given the power of the biological process of evolution to adapt to ever, changing environments, it is surprising that the number of studies applying and explicitly studying their artificial counterpart of GP in dynamic environments have been minimal (Dempsey 2007). Despite the existence of a recent Special Issue in the GP Journal on Dynamic environments (Yang et al 2006), none of the four articles actually dealt with GP directly (e.g., Wang and Wineberg (2006)). While some applications in dynamic environments have been undertaken in the past two years (e.g., Wagner et al 2007, Hansen et al 2007, Jakobović and Budin 2006, and Kibria and Li 2006, there has been little analysis of the behavior of GP in these environments.…”