Survey on Multi-Agent Q-Learning frameworks for resource management in wireless sensor network

Abstract

This report aims to survey multi-agent Q-Learning algorithms, analyze different game theory frameworks used, address each framework's applications, and report challenges and future directions. The target application for this study is resource management in the wireless sensor network.In the first section, the author provided an introduction regarding the applications of wireless sensor networks. After that, the author presented a summary of the Q-Learning algorithm, a well-known classic solution for model-free reinforcement learning problems.In the third section, the author extended the Q-Learning algorithm for multi-agent scenarios and discussed its challenges.In the fourth section, the author surveyed sets of game-theoretic frameworks that researchers used to address this problem for resource allocation and task scheduling in the wireless sensor networks. Lastly, the author mentioned some interesting open challenges in this domain.Wireless sensor network provides online monitoring capabilities in situations that are not accessible (for example: controlling the temperature of a nuclear reactor, or invasive brain, or muscular signal monitoring).Usually, Wireless sensor nodes are heterogeneous, energy-constrained, and tend to operate in dynamic and unclear situations. In these situations, nodes need to learn how to cooperate over tasks and resources (including power and bandwidth). It implies that we design a framework that allows wireless sensor nodes to adapt to the new situation. In these scenarios, Reinforcement Learning is an immeasurable solution. [8].Recently, reinforcement learning (RL) becomes a trend in various autonomous decision-making tasks, whether sequential or simultaneous. For example, tasks related to solving strategic games, or sensor and communication networks, finances, social science, etc. [12].