Aggregator election in wireless sensor networks: A distributed reinforcement learning approach

Abstract

Nowadays, artificial intelligence techniques are used in various fields of wireless sensor networks. Due to resource constraints in these types of networks, many studies focus on minimizing energy consumption and increasing the lifetime of the networks. Data aggregation is a powerful technique that it reduces the energy consumption in the network. In this paper, we've provided a distributed approach based on reinforcement learning and using learning automata for solving the problem of selection of aggregator in wireless sensor networks. We compared our method with DRLR and ECHSSDA algorithms. The results show that the proposed method significantly reduces energy consumption in DRLR and outperforms ECHSSDA, especially when the environment has low density.