Building Norms-Adaptable Agents from Potential Norms Detection Technique (PNDT)

Abstract

This paper presents a contribution to research on norms detection by proposing a technique, which is called the Potential Norms Detection Technique (PNDT). The literature proposes that an agent changes or updates its norms based on the variables of the local environment and the amount of thinking about its behaviour. Consequently, any changes on these two variables cause the agent to use the PNDT to update the norms in complying with the domain’s normative protocol. This technique enables an agent to update its norms even in the absence of sanctions from a third-party enforcement authority as found in some work, which entail sanctions by a third-party to detect and identify the norms. The PNDT consists of five components: agent’s belief base; observation process; Potential Norms Mining Algorithm (PNMA) to detect the potential norms and identify the normative protocol; verification process, which verifies the detected potential norms; and updating process, which updates the agent’s belief base with new normative protocol. The authors then demonstrate the operation of the algorithm by testing it on a typical scenario and analyse the results on several issues.

A proposed norms' benefits awareness framework for norms adoption

“…The phenomenon of imitation has been described as "When in Rome, do as the Romans do". In a similar work, Mahmoud et al [12] presented a self-enforcing agent which detects norms based on beliefs that are triggered by emotions of imitating the majority and the agent extracts the norms using a norms mining algorithm. The mechanisms demonstrate advanced behavior whereby an agent observes and replicates another's behavior regardless of the benefit gained.…”

“…Research on norms identification recently appeared in the literature expounding issues and techniques on this research domain [8,9,12,17]. For example, Savarimuthu et al [9] proposed a norm identification technique in which a visitor agent infers the norms of an agents' community without the norms being explicitly given to the agent.…”

“…These mechanisms are based on sanction [3,4], imitation [5,6,7], or social learning [7,8]. Consequently, research on norms identification and detection have recently appeared in the literature expounding issues and techniques on this research domain [8,12,18,19,20]. For example, Savarimuthu et al [9] proposed a norm identification technique in which a visitor agent infers the norm based on a sanctioned agent and after inferring, the agent adopts the norm.…”

“…In a similar work, Mahmoud et al [12] presented a self-enforcing agent and Sen and Airiau [13] proposed a social learning theory.…”

Norm-Based Behavior Regulating Technique for Multi-Agent in Complex Adaptive Systems

“…Generally, Norms are agreed upon rules set by population indicating expected actions to perform that are either prohibitive, obligatory or permissive [4]- [6]. However, to persuade agents to comply with norms, normative systems should be able to extend rewards or penalties [7]- [9].…”