Improving Router Cooperation in Mobile Wireless Sensor Networks Using Reinforcement Learning

This paper proposes to promote cooperative routing for homogeneous mobile wireless sensor networks (mWSNs) using a scalable, distributed incentive-based mechanism with reasonable resource requirements using reinforcement learning (RL). In particular, Q-learning which is a well-known RL method was integrated an existing Continuous Value Cooperation Protocol (CVCP). We also studied their effects on the efficiency in non-cooperative mWSNs and propose a good routing strategy under constrained conditions such as network traffic load, degree of mobility and path loss exponent.