Towards motivation-based decisions for worth goals

Abstract

In this paper we present a motivational mechanism to generate and determine the worth of goals and to represent various constraints involved in satisfying a goal. The work builds on the SMART agent framework and adds to the growing body of work that is attempting to extend the abilities of autonomous agents past the constraints of the traditional symbolic approaches to AI. The paper represents a first step in increasing an agent's autonomy in the domain of e-commerce, specifically enabling the agent to dynamically set issue parameters in relation to the importance of the issue and the effects of any existing constraints.