A context-aware approach of multiple criteria decision making for social network analysis

The context-awareness is the knowledge background that the agents take into account in doing their decisions. It allows to consider the multidimensionality of the interaction space where decisions and criteria evolve. The idea is to redesign the concept of space making it smart and allowing the single agents to communicate and interact through the global social environment. In a social network, the interaction among nodes, or agents, is able to carry on a strong cognitive value. This process has a strong impact on the context, in which every agent operates. A network node becomes an aware agent capable of interacting within a community or among different communities, sharing its own knowledge, perception and awareness. Through social interaction, the network structure, promoting transmission of social objects, enclosing opinions, data, influences, decisions, enhances the collective knowledge and individual awareness. The space becomes a powerful environment, where each phenomenon has its own dynamic and moves as a flow across bubbles. The intensity of the flow depends on how each agent has a cognition of its neighborhood, of the environment in which it operates and, therefore, of the entire space in which it can act. This bio-inspired and socialbased process rules the situated cognition of each agent inside a network smart space. In this perspective Multiple Criteria Decision Making (MCDM) has to take into consideration the evolution of the preferences due to interaction with other agents within a smart space. This approach constitutes a
change of paradigm for MCDM and requires a re-conceptualization of the main
aspects in the discipline. Each agent takes decisions considering not only its current individual preferences, but also the preferences arising from the smart space, for example preferences of agents with which there is an interaction. This increases the complexity of MCDM models that have to handle also the greater uncertainty of the considered environment. We discuss some implications of considering a smart space within current MCDM models.