2002 | OriginalPaper | Buchkapitel
Preliminaries
verfasst von : Dirk V. Arnold
Erschienen in: Noisy Optimization With Evolution Strategies
Verlag: Springer US
Aktivieren Sie unsere intelligente Suche um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
In this chapter, we define the strategies and fitness environments to be analyzed in the remainder of this work. It serves the purpose of introducing basic definitions and terminology, and of preparing the reader for the analyses presented in the following chapters. We do not attempt to introduce evolution strategies in their full generality, but rather restrict ourselves to real-valued search spaces with fitness functions f: ℝNℝ and to $$ ({\mu \mathord{\left/ {\vphantom {\mu {\rho \mathop + \limits_, \lambda }}} \right. \kern-\nulldelimiterspace} {\rho \mathop + \limits_, \lambda }}) - ES $$ with intermediate recombination and with isotropic normal mutations. This choice of strategy is motivated both by that is relatively amenable to mathematical analysis and by its pervasiveness in the literature. References to other strategy variants are included where appropriate.