Specialization in Collective Behavior Systems : The Role of Emergent Specialization in Collective Behavior

ISBN-13: 9783838322506 / Angielski / Miękka / 2009 / 188 str.

Geoff Nitschke

Opis

Specialization in Collective Behavior Systems describes a novel method called: Collective Neuro-Evolution (CONE) that operates within computer simulation to solve collective behavior problems. A collective behavior problem is a task that can only be solved by multiple agents (computer programs) working cooperatively. The book s topic is situated in the neuro-evolution research field: the intersection of evolutionary computation, and neural computation research. Given a collective behavior problem, CONE designs multi-agent behavior such that agents cooperate in order to solve the problem. CONE works via adapting agent behavior and agent interactions during a computer simulation, and purposefully facilitates and uses specialization that emerges as agents interact. Using emergent specialization as a problem solving mechanism allows CONE to outperform related methods that also solve computer simulated collective behavior problems.

O autorze

Geoff Nitschke is a post-doctoral research fellow at the Computational Intelligence Research Laboratory, at the University of Pretoria, South Africa. He received his doctorate in Computer Science from the Vrije Universiteit of Amsterdam, The Netherlands. His research focus includes artificial life, neuro-evolution, and collective behavior.