Autonomousagentsin the real world must be capableof asynchronous goal generation. However, one consequence of this ability is that the agent may generate a substantial number of goals, but only a small number of these will be relevant at any one time. Therefore, there is a need for tractable mechanisms to manage a changing and potentially large number of goals. This paper presents both a framework for the design of agents that have the capability to generate and manage their own top level goals, “motivated agency”, and an implementation of part of this agent architecture, “alarms”. The alarm-processing machinery serves to focus the attention of an agent on a limited number of the most salient goals regardless of the number of possible goals that the agent can pursue or their distribution in time. In this way, a resource-bounded autonomous agent can employ modern planning methods to greater effect.
Timothy J. Norman, Derek Long