Aspects of Active Norm Learning and the Effect of Lying on Norm Emergence in Agent Societies

Coordination, Organizations, Institutions, and Norms in Agent Systems X

Leeke

et al. 2015

Abstract. Conventions are an important concept in multi-agent systems as they allow increased coordination amongst agents and hence a more efficient system. Encouraging and directing convention emergence has been the focus of much research, particularly through the use of fixed strategy agents. In this paper we apply temporary interventions using fixed strategy agents to destabilise an established convention by (i) replacing it with another convention of our choosing, and (ii) allowing it to destabilise in such a way that no other convention explicitly replaces it. We show that these interventions are effective and investigate the minimum level of intervention needed.

Section: Related Workmentioning

confidence: 99%

Destabilising Conventions Using Temporary Interventions

Coordination, Organizations, Institutions, and Norms in Agent Systems X

Leeke

et al. 2015

“…Conventions place 'soft constraints' on agent choices by encouraging mutually beneficial behaviour by adherence to the convention. Unlike norms there is no explicit punishment for going against the convention but doing so is likely to incur a cost to the agent due to increased clashes often represented as a negative interaction payoff [10,18]. Conventions can thus be described as "an equilibrium everyone expects in interactions that have more than one equilibrium" [24].…”

Section: Related Workmentioning

confidence: 99%

Convention Emergence in Partially Observable Topologies

Autonomous Agents and Multiagent Systems

2017

Abstract. In multi-agent systems it is often desirable for agents to adhere to standards of behaviour that minimise clashes and wasting of (limited) resources. In situations where it is not possible or desirable to dictate these standards globally or via centralised control, convention emergence offers a lightweight and rapid alternative. Placing fixed strategy agents within a population, whose interactions are constrained by an underlying network, has been shown to facilitate faster convention emergence with some degree of control. Placing these fixed strategy agents at topologically influential locations (such as high-degree nodes) increases their effectiveness. However, finding such influential locations often assumes that the whole network is visible or that it is feasible to inspect the whole network in a computationally practical time, a fact not guaranteed in many real-world scenarios. We present an algorithm, PO-Place, that finds influential nodes given a finite number of network observations. We show that PO-Place finds sets of nodes with similar reach and influence to the set of high-degree nodes and we then compare the performance of PO-Place to degree placement for convention emergence in several real-world topologies.

“…Norms typically imply an obligation or prohibition on agents with regards to a specific action. Failure to adhere to norms and exhibit the expected behaviour is often associated with punishments or sanctions [1,3,11,19]. Alternatively, agents may be explicitly rewarded for adherence to norms.…”

Section: Related Workmentioning

confidence: 99%

Manipulating Conventions in a Particle-Based Topology

Lecture Notes in Computer Science

2016

Abstract. Coordination is essential to the effective operation of multiagent systems. Convention emergence offers a low-cost and decentralised method of ensuring compatible actions and behaviour, without requiring the imposition of global rules. This is of particular importance in environments with no centralised control or where agents belong to different, possibly conflicting, parties. The timely emergence of robust conventions can be facilitated and manipulated via the use of fixed strategy agents, who attempt to influence others into adopting a particular strategy. Although fixed strategy agents have previously been investigated, they have not been considered in dynamic networks. In this paper, we explore the emergence of conventions within a dynamic network, and examine the effectiveness of fixed strategy agents in this context. Using established placement heuristics we show how such agents can encourage convention emergence, and we examine the impact of the dynamic nature of the network. We introduce a new heuristic, Life-Degree, to enable this investigation. Finally, we consider the ability of fixed strategy agents to manipulate already established conventions, and investigate the effectiveness of placement heuristics in this domain.