Mechanisms to Restrict Exploitation and Improve Societal Performance in Multi-Agent Systems

Mechanisms to Restrict Exploitation and Improve Societal Performance in Multi-Agent Systems

Sharmila Savarimuthu (University of Otago, New Zealand), Martin Purvis (University of Otago, New Zealand), Maryam Purvis (University of Otago, New Zealand) and Mariusz Nowostawski (University of Otago, New Zealand)
Copyright: © 2009 |Pages: 13
DOI: 10.4018/978-1-59904-576-4.ch011


Societies are made of different kinds of agents, some cooperative and uncooperative. Uncooperative agents tend to reduce the overall performance of the society, due to exploitation practices. In the real world, it is not possible to decimate all the uncooperative agents; thus the objective of this research is to design and implement mechanisms that will improve the overall benefit of the society without excluding uncooperative agents. The mechanisms that we have designed include referrals and resource restrictions. A referral scheme is used to identify and distinguish noncooperators and cooperators. Resource restriction mechanisms are used to restrict noncooperators from selfish resource utilization. Experimental results are presented describing how these mechanisms operate.
Chapter Preview

Cooperative Behaviour In Multi-Agent Society

For a society to operate effectively, agents within the society must obey certain social rules and norms. So far, much of the focus in this area has been on work devoted to the identification of malevolent agents, where the goal is to identify a noncooperator and exclude it from the society. However, in the real world, it is not going to be applicable in all situations. Our focus is on situations where society members are behaving in an uncooperative manner, but are not necessarily “evil” and deserving of expulsion. This is the issue of the “Tragedy of the Commons” (Hardin, 1968).

Tragedy of the Commons

In Hardin’s classic paper (Hardin, 1968), “Tragedy of the Commons,” he outlines the “tragedy.” A common pasture is open to herders, each of which tries to maintain as many cattle as possible on the commons. A herder will reckon that the positive benefits of adding one additional animal will all go to him, alone, whereas the negative effects from overgrazing of that one additional animal will be shared borne by all the herders. Accordingly, self-interested herders may continue adding one more animal to their herds, even if they know that collectively this is destroying the commons. The question is: how are restrict selfish herders to avoid the tragedy?

The Tragedy of the Commons can be related to the “Prisoner’s Dilemma” situation (Axelrod, 1984). Two collaborating criminals are imprisoned and questioned separately. Each criminal may cooperate with his fellow criminal by refusing to divulge details of the crime or defect by ratting on his colleague. It is possible to establish a reward structure (see Figure 1) such that:

  • If both criminals cooperate they get a reward, R,

  • If they both defect, they are punished (punishment, P),

  • If one player defects and the other cooperates, then the defector gets high reward (temptation, T) and the other gets a severe punishment (sucker, S)

  • And T > R > P > S, and 2R > T + S

Under these reward conditions, each individual criminal will reason that if the other

  • Cooperates, he does better by defecting, and if the other

  • Defects, he also does better by defecting.

    Figure 1.

    Payoff matrix for prisoner’s dilemma

Thus, the Nash equilibrium situation for this game is for both players to defect, even though they would collectively get a higher reward if they were both to cooperate. The Tragedy of the Commons can be likened to a situation in which the individual herder is playing the Prisoner’s Dilemma game against the collection of all the other herders: his selfish interests lead him to defect, even though they are all better off if they cooperate.

Another cooperation game that is discussed in the literature is the Stag Hunt game. The metaphor here is two hunters who may cooperate to hunt a stag (high reward, S). If they operate by themselves, they each can only catch a rabbit (lower reward, R). A hunter seeking to hunt a stag without cooperation gets nothing. But there is no sucker’s reward here. The reward structure is shown below (see Figure 2).

Figure 2.

Payoff matrix for Stag Hunt

Complete Chapter List

Search this Book: