Towards a Framework for Detecting Opportunism in Multi-Agent Systems

摘要

Consider a common social scenario. A seller sells a cup to a buyer and it is known by the seller beforehand that the cup is actually broken. The buyer buys the cup without knowing it is broken. Since the buyer's value gets demoted, the behavior performed by the seller is usually forbidden by social norms. Such a social behavior intentionally performed by the seller is first named opportunistic behavior (or opportunism) by economist Williamson [6]. It is a typical social behavior that is motivated by self-interest and takes advantage of knowledge asymmetry about the behavior to achieve own gains, regardless of the principles [3]. This definition implies that, given a social context, opportunistic behavior results in promoting agents' own value while demoting social value. Therefore, it is prohibited by norms in most societies. In the context of multi-agent systems, we constrain such a selfish behavior through setting enforcement norms, in the sense that agents receive a corresponding sanction when they violate the norm. On the one hand, it is important to detect it, as it has undesirable results for the participating agents. On the other hand, as opportunism is always in the form of cheating, deception and betrayal, meaning that the system does not know what the agent performs or even the motivation behind it (for example, in a distributed system), monitors can only observe agents' opportunistic behavior indirectly. Therefore, there has to be a monitoring mechanism that can detect the performance of opportunistic behavior in the system. This paper introduces the idea of using a logical framework based on the specification of actions to verify whether agents in the system perform opportunistic behavior.