|
|||
A Population-Differential Method of Monitoring Success and Failure in CoevolutionAri Bader-Natal and Jordan B. Pollack DEMO Lab, Computer Science Department MS018, Brandeis University, Waltham, MA 02454 USAari@cs.brandeis.edu pollack@cs.brandeis.edu Abstract. Coevolutionary algorithms require no domain-specific measure of objective fitness, enabling these these algorithms to be applied to domains for which no objective metric is known or for which known metrics are too expensive. But this flexibility comes at the expense of accountabilitiy. Past work on monitoring has focused on measuring success, but has not been able to provide feedback on failure. This limitation is due to a common reliance on “best-of-generation” (BOG) based analysis [2], and we propose a population-differential analysis based on an alternate “all-of-generation” (AOG) framework that is not similarly limited. LNCS 3102, p. 585 f. lncs@springer.de
|