Their Problems are My Problems

Abstract

In Multi-Agent Systems there has been quite an amount of research on how to model conflicts in (intelligent) agent architectures for example by imitating mental states and augmenting them with social notions. Nevertheless, a great deal of investigation is missing on the transition between internal and external conflicts. Arguments in favor of taking a closer look at this transition include a practical one from applications where internal conflicts may easily become external conflicts and vice versa, and a technical one motivated by the existence of more efficient algorithms for solving internal conflicts.

In this chapter we will use a rather technical and restricted notion of “conflict” to exploit the mathematical model of distributed constraint satisfaction problems that covers internal as well as external conflicts in a unique framework. In this model conflicts can be identified as “inconsistencies” and there are several algorithms for conflict detection, conflict avoidance and conflict resolution. After introducing models for internal and external conflicts and algorithms to solve them, we will focus on a new concept called AURECON (autonomous dynamic reconfiguration) that provides a means for balancing on the transition between internal and external conflict.

This chapter summarizes the results presented in (Hannebauer, 2000a) and (Hannebauer, 2000b) and extends them in several ways.