International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 27 - Number 6 |
Year of Publication: 2011 |
Authors: B. Sasikumar, V. Vasudevan |
10.5120/3301-4509 |
B. Sasikumar, V. Vasudevan . Multi Agent System based TCP for Wireless Networks. International Journal of Computer Applications. 27, 6 ( August 2011), 45-50. DOI=10.5120/3301-4509
Due to the rapid evolution of networking and electronic capabilities, the use of wireless devices is becoming ubiquitous. Most of the wireless networks have fixed infrastructure and requires higher data rates. Due to high data rate in wireless networks the Transmission Control Protocol (TCP) performance has a broad and significant impact on data applications, and is essential towards massive deployment of service-providing agents and their widespread social acceptance. Wireless link losses result in poor TCP throughput since losses are professed as congestion by TCP. In order to increase the throughput, a Multi agent System (MAS) named Reusable Environment for Task-Structured Intelligent Networked Agents (RETSINA) has been introduced. This agent has the capability to reduce the impact of losses on TCP throughput and latency. In this work, we considered a video game application in which a networked virtual environment with high fidelity sound and video is created. Moreover, to efficiently use the video game application in a wireless environment, an Unreal Tournament Semi-Automated Force (UTSAF) multi agent has been proposed. UTSAF is a middleware written to take advantage of the power of gaming application systems by allowing them to participate in distributed simulations. The personal agent communicates wirelessly, in order to find the most appropriate option to serve the user. In addition, we comprehensively evaluated the performance of TCP with and without these multi agent platforms as technology choice and investigated its efficiency in a number of cases. We find that the simulations with the multi agent can reach the TCP throughput by as much as 95 % and the latency get reduced by 1-2 % as the simulation time steps in for every 100 seconds.