Wireless Sensor Network Topology Control Based on Agent

Abstract

Based on multi-Agent and learning-reinforcement adaptive topology control algorithm of wireless sensor network, this paper abstracts the wireless sensor network topology control into multi-Agent and global-coordination optimization problem on the basis of payment and localizes the global evaluation function by taking advantage of SparseQ in order to form effective topological structure and optimized data forwarding path in random deployment mode. This paper also probes into forming initial topological structure and data forwarding path by the interaction of detection information and return information among local Agents, and probes into ensuring topology's continual updating and optimization by the renovation of return information and path selection evaluation values in the process of data forwarding. The simulation shows that the algorithm improves the data transmission quality and energy efficiency in the specific application of wireless sensor network in random deployment mode.

Keywords

• Wireless sensor networks
• Agent
• Mobile agent
• Multi-agent system
• Topology control
• Multi-agent reinforcement learning.