A novel mobile charging planning method based on swarm reinforcement learning in wireless sensor networks

Abstract

In order to solve the problem of energy supplement in large-scale wireless sensor networks (WSNs), this paper investigates the charging planning problem by introduced multiple wireless charger equipment (WCE). We first established the optimisation model of the multi-WCE charging planning problem to minimise the total charging time and the total energy consumption of the WCE. Then, the problem is modelled as a reinforcement learning process, and the time step, state space, action space, state transfer function and reward function are designed. Moreover, based on the idea of swarm intelligence optimisation method, a multi-learners' strategy is introduced to enable multi-learners to parallel learning, so as to accelerate the solution finding speed. Therefore, a discrete firework Q-learning algorithm is proposed to solve the problem. Experiments show that the proposed algorithm outperforms the baseline algorithms in different network scales.