Mechanisms for norm emergence and norm identification in multi-agent societies

Norms are standards of behaviour expected of the members of a society. Norms in human societies help in sustaining social order and facilitating cooperation and coordination among agents. Researchers in multi-agent systems investigate how the concept of norms can serve a similar purpose in artificial societies with autonomous agents. This thesis contributes to two aspects of the study of norms in multi-agent systems through the investigation of mechanisms for norm emergence and norm identification. With the advent of digital societies such as Second Life, software agents that reside in these societies are expected to abide by the norms of those societies. Most works on norms in multi-agent systems assume that agents know the norms a priori. Though this is important, norms that are not explicitly specified by the designers of the society may emerge in open agent societies. Thus there is a need for the study of mechanisms for artificial agent societies which can facilitate norm emergence based on interactions between agents. To that end the first part of this thesis describes a role model mechanism for norm emergence. The thesis also describes how norms can emerge in connection with different types of network topologies. A particle-collision model for constructing dynamic network topologies has been applied to model how two societies can be brought together. Using such a model, norm emergence on dynamic network topologies have been studied. With the uptake of virtual environments which are open and dynamic, agents residing in these societies should be endowed with mechanisms that facilitate norm identification. To that end, the second part of the thesis investigates how a software agent comes to know the norms of the society that it is a part of. To achieve