How to do the Right Thing

TLDR

An agent is viewed as a collection of competence modules. This paper presents a novel approach to action selection for an autonomous agent and introduces a concrete algorithm with a detailed results account. Action selection is modelled as an emergent property of activation/inhibition dynamics among competence modules, combining planner and reactive system traits, and offers global parameters to tune goal orientation, adaptivity, conflict sensitivity, and speed. The presented algorithm yields detailed results demonstrating its effectiveness in action selection. Keywords: control architectures for autonomous agents, action selection algorithms, spreading activation networks.

Abstract

Abstract This paper presents a novel approach to the problem of action selection for an autonomous agent. An agent is viewed as a collection of competence modules. Action selection is modelled as an emergent property of an activation/inhibition dynamics among these modules. A concrete action selection algorithm is presented and a detailed account of the results is given. This algorithm combines characteristics of both traditional planners and reactive systems. It provides global parameters, which one can use to tune the action selection behavior along several criteria, such as goal orientedness versus situation orientedness, bias towards ongoing plans versus adaptivity, and sensitivity to goal conflicts and 'thoughtfulness' versus speed. KEYWORDS: Control architectures for autonomous agentsaction selection algorithmsspreading activation networks.