A Q-Routing Protocol Using Self-Aware Approach for Mobile Ad hoc Networks | Science Publications

American Journal of Applied Sciences

Miznah Al-Rodhaan¹  Amal Alharbi¹  Abdullah Al-Dhalaan¹ 

关键词: Cognitive Packet Network (CPN);    Mobile Ad hoc Network (MANET);    Q-Routing;    Reinforcement Learning (RL);    Self-Aware Networks (SAN);   

DOI  :  10.3844/ajassp.2015.1014.1022

学科分类：自然科学（综合）

来源: Science Publications

【 摘 要 】

Mobile Ad hoc Networks (MANET) are self-organized networks thatare characterized by dynamic topologies in time and space. This creates an instableenvironment, where classical routing approaches cannot achieve highperformance. Thus, adaptive routing is necessary to handle the challenges inMANETs. Furthermore, it is necessary for nodes to be self-aware i.e., able todiscover neighbors, links and paths when needed. This paper proposes a newadaptive Mobile Ad hoc Networks (MANET) routing algorithm to find and maintainpaths that provide the needed Quality of Service (QoS) for network traffic usinga low-complexity bio-inspired learning paradigm. It combines the self-awareapproach in Cognitive Packets Network (CPN) with a Q-routing inspired pathselection mechanism. CPN is a distributed adaptive routing protocol that usesthree types of packets: Smart Packets for route discovery, Data Packets forcarrying data payload and Acknowledgments to bring back feedback informationfor the Reinforcement Learning rewardfunction. The research defines a Q-routing reward function as a combination ofhigh stability and low delay path criteria to discover long-lived routeswithout disrupting the overall delay. The algorithm uses Acknowledgment-basedfeedback for Q-routing to make routing decisions that adapt on line to networkchanges allowing nodes to learn efficient routing policies. Simulation Resultsshow how the reward function handles the network changing topology to selectpaths that improve QoS delivered.

【 授权许可】

Unknown