Amol Dattatraya Mail and Amitabha Mukerjee
This research focuses on a specific class of agents called "situated agents", which use minimal communication and rely mostly on changes in the environment as their cue for action. Some early successes of this model, especially in robotics, have led to an intense debate over this class of models as a whole. One of the issues on which attention has been drawn is that of conflicts between such agents. In this work we investigate a cyclic conflict that results in infinite looping between agents and has a severe debilitating effect on performance. We present some new results in the debate, and compare this problem with similar cyclicity observed in planning systems, meta-level planners, distributed agent models and hybrid situated models. Some of our results are - 1. The probability of such cycles developing increases as the situated agents become more useful (phase transition). 2. Control methods for avoiding cycles such as prioritization are unreliable. 3. Behavior refinement methods that reliably avoid these conflicts (either by refining the stimulus, or by weakening the action) lead to weaker functionality. 4. Conflicts place a bound on the power of new behavior that can be added to a situated system. We use these and some other results to examine the assumption of modularity in situated agents. Viewing chess playing as a situated activity, we discuss how our results hold relevance to the design of chess players.