A meritocratic trust-based group formation in an IoT environment for smart cities

• Published in: Future generation computer systems 2020-07, Vol.108, p.34-45
• Main Authors: Fortino, G., Fotia, L., Messina, F., Rosaci, D., Sarné, G.M.L.
• Format: Article
• Language: English
• Subjects: Competitive agents; Group; Internet-of-Things; Meritocracy; Nash equilibrium; Smart city; Trust
• ISSN: 0167-739X; 1872-7115
• DOI: 10.1016/j.future.2020.02.035

Smart cities are built on top of heterogeneous IoT infrastructures, that can be viewed as communities of software agents (the intelligent objects) that interact with each other to realize complex activities. These agents operate on behalf of users that need services; for these reason agents are often in competition with each other. On the other hand, an agent can often benefit from collaborating with other agents in some circumstances, exchanging information and services. Under this viewpoint, the task of finding the best partners to collaborate is a key task for an agent. A general consensus exists about the benefits deriving by forming friendships and groups for mutual cooperation inside competitive Multi-Agent Systems (MASs). In this respect, the existing proposals are usually addressed to maximize the profit at the level of individual agent or group. Unfortunately, the most part of these approaches could advantage the most aggressive agents, also in presence of bad social behaviors. This is not a desired scenario in a smart city environment. A possible solution to this problem is that of promoting correct behaviors and meritocracy inside agent communities. To this aim, we propose to model the competitive MAS scenario in the framework of non cooperative games by assuming to represent (i) the trustworthiness of agents relationships by means of a trust model and (ii) the capability of a community to provide its members with a good environment by means of its social capital. As a result, a group formation algorithm capable to asymptotically maximize the social capital is proposed. This algorithm highlights two main features: (i) the computed solution is a Nash equilibrium in the considered game and (ii) the only rewarded agents are those having the most correct behaviors. •Promoting meritocracy inside agent communities in smart cities.•A Multi Agent System (MAS) modeled in the framework of non cooperative games.•Exploiting trustworthiness of agents and social capital.•Design of a group formation algorithm to maximize the social capital.•The solution computed by the algorithm is a Nash equilibrium.