No abstract

No abstract

Device-to-device (D2D) communication is an essential feature for the future cellular networks as it increases spectrum efficiency by reusing resources between cellular and D2D users. However, the performance of the overall system can degrade if there is no proper control over interferences produced by the D2D users. Efficient resource allocation among D2D User equipments (UE) in a cellular network is desirable since it helps to provide a suitable interference management system. In this paper, we propose a cooperative reinforcement learning algorithm for adaptive resource allocation, which contributes to improving system throughput. In order to avoid selfish devices, which try to increase the throughput independently, we consider cooperation between devices as promising approach to significantly improve the overall system throughput. We impose cooperation by sharing the value function/learned policies between devices and incorporating a neighboring factor. We incorporate the set of states with the appropriate number of system-defined variables, which increases the observation space and consequently improves the accuracy of the learning algorithm. Finally, we compare our work with existing distributed reinforcement learning and random allocation of resources. Simulation results show that the proposed resource allocation algorithm outperforms both existing methods while varying the number of D2D users and transmission power in terms of overall system throughput, as well as D2D throughput by proper Resource block (RB)-power level combination with fairness measure and improving the Quality of service (QoS) by efficient controlling of the interference level.

No abstract

Wireless sensor networks (WSN) are an attractive platform for cyber physical systems. A typical WSN application is composed of different tasks which need to be scheduled on each sensor node. However, the severe energy limitations pose a particular challenge for developing WSN applications, and the scheduling of tasks has typically a strong influence on the achievable performance and energy consumption. In this paper we propose a method for scheduling the tasks using cooperative reinforcement learning (RL) where each node determines the next task based on the observed application behavior. In this RL framework we can trade the application performance and the required energy consumption by a weighted reward function and can therefore achieve different energy/performance results of the overall application. By exchanging data among neighboring nodes we can further improve this energy/performance tradeoff. We evaluate our approach in an target tracking application. Our simulations show that cooperative approaches are superior to non-cooperative approaches for this kind of applications.