A Game-Theoretic Approach for Cost-Effective Multicast Routing in the Internet of Things

Internet of Things (IoT) devices have enabled communications in resource-limited computing environments. Sensor nodes from the multiple IoT devices collectively work for many applications, such as disaster management, border security management, smart farming, smart cities, etc. In such applications, the data from a single source node is often destined for multiple nodes. Multicast communication is preferred over unicast or broadcast communication for such applications, as multicast uses fewer resources. Efficient construction of the multicast tree leads to cost-effective multicast transmission. This article introduces a path selection game (PSGame), a game-theoretic approach that formulates the construction problem of the least-cost multicast tree as a potential game. Our proposed path selection algorithm (PSA) quickly converges to the pure Nash equilibrium (PNE), bringing the least cost multicast tree. Our findings show that the overhead incurred in terms of energy consumption and delay is minimal in the proposed algorithm compared to other mechanisms. The theoretical analysis proves that the proposed algorithm quickly converges to PNE in $O(n.r_{\max })$ steps. It also proves that the cost ratio between the proposed solution and the centralized optimum will be bounded by $\log (n)$ . The numerical analysis substantiates the theoretical analysis.

I. Introduction

Internet of Things (IoT) devices are emerging rapidly with the Quality of Service (QoS) guarantees. Sensors are used in IoT devices to collect real-time data from the physical world, and this data is transmitted to various devices via the Internet. Fig. 1 depicts a smart city scenario where different sensor nodes collect and transmit the data to the control server via base stations. The control server process this data and provides the required information to different destinations (fire stations, police stations, traffic control rooms, gas stations, etc.). According to IoT architecture [1], trillions of things embedded with sensors are deployed in the data sensing layer to sense the data from the environment. These things are connected in a wired or wireless fashion by various kinds of network infrastructure [2]. Fig. 1.

Example of a smart city.