Bayesian coalition game for the internet of things: an ambient intelligence-based evaluation

With the evolution of the Internet and related technologies, there has been an evolution of new paradigm, which is the Internet of Things, IoT. In the IoT, a large number of objects/devices on the Internet are connected to one another for information sharing, irrespective of their locations. These devices may be interconnected with one another using various network protocols and standards to exchange information between them. The underlying network used for information exchange generally has built-in intelligence, which is called ambient intelligence, so that it can make adaptive decisions for information exchange between these objects in theh IoT. This article provides a performance evaluation of the Bayesian coalition game among these objects in the IoT environment by using the concepts of game theory and LA. In comparison to the existing solutions, LA are assumed to be the players in the game having variable learning rates in the coalition game. Most of the existing solutions have considered constant learning rates of the players in the game, which may lead to the possibility of local optima at some points. Each player decides its actions using competitive learning, having variable learning rates, based on the newly defined utility function, which leads to the achievment of a Nash equilibrium in the game quickly. Each player receives feedback from the environment corresponding to the actions taken in a unit interval of time. The performance of the proposed scheme was evaluated with respect to various performance evaluation metrics. The results obtained show that the proposed scheme is useful in the IoT environment.