Multi-Hop Wireless Network Optimization Solution using Automatic Distributed Joint Routing for Delay Sensitive Applications
I S. Akila Rajini, J. Raja Shahul Hameed · 2014
Multi-hop wireless network can be modeled as a dynamic network, consisting of several interconnected wireless nodes, which aim to jointly optimize the overall network utility, given the resource constraints of the wireless communication channels and also, importantly, the mutual interferences (coupling) resulting when nodes are simultaneously transmitting. The proposed system provides a distributed routing and power control algorithm that enables nodes in a multi-hop network to autonomously optimize the overall performance of delay-sensitive applications by determining their routing and transmission power to maximize the network utility, in a dynamic environment. This is achieved with the help of Markov Decision Process with distributed computation of the optimal policy, which enables individual nodes to make optimal decisions for their cross-layer strategies autonomously, by relying only on their local available information rather than acquiring global information which would cause a large delay and high communication overhead. The reinforcement-learning method is used to find the optimized policy when the dynamics are unknown. A pricing-based distributed resource allocation framework is also adopted in proposed approach. By using this cooperative approach, the video quality can be improved.