ABSTRACT
We investigate minimally cognitive agents for attention and control of action. A visual agent (VA) can attend to one target while ignoring another and can then reallocate processing. Here we review some of our previous results with VA, showing how an analysis of a minimally cognitive agent can both support previous ideas of selective attention and action, as well as suggest new possibilities. We investigate a simple case in which multiple agents are combined to solve tasks that no single agent can complete on its own.
