Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
485657 | Procedia Computer Science | 2015 | 7 Pages |
Wireless Wireless network finds application in military environments, emergency, rescue operations and medical monitoring due to its self-configuring nature. As the availability of resources such as processing power, buffer capacity and energy are limited in wireless networks; it is required to devise efficient algorithms for packet forwarding. Due to the dynamic nature of the wireless environment, the traditional packet forwarding strategies cannot guarantee good network performance every time. This paper proposes a method for learning data flow rates in wireless network to improve quality of service in the network. Each node in the network learns the environment using reinforcement learning approach and selects appropriate neighbours for packet forwarding. In order to improve the learning capacity of nodes, the hierarchical docition technique is employed. Docition applied to each layer of network, which selects a set of special nodes which has more information about the environment and share this information with less informative nodes. The algorithm is tested in a geographical routing protocol and the results indicate improved network performance.