Models for strategy choice in random or potentially deceptive environments

Abstract

Two simulation models are presented illustrative of decision and behavioural choice issues for complex adaptive systems involving cognitive agents. The first illustrates the value of managed instability for complex systems. It explores conditions under which a system able maintain itself in an unstable state gains an advantage in behavioural flexibility offsetting the extra expense involved in maintaining that state against instability. The second model considers the trade-off between quickness and accuracy of response in risk/reward identification, in order to differentiate conditions under which rapid or considered responses are favoured. The second model is currently being used to distinguish cases where cognitive agents are preferable to non-cognitive agents and vice versa.