Optimal Itinerary Analysis for Mobile Agents in Ad Hoc Wireless Sensor Networks

One of the most important problems studied in any sensor network is data fusion. Client/server paradigm has been a commonly used computing model in traditional distributed sensor networks (DSNs). However, the deployment of wireless sensor networks (WSNs) and its ad hoc nature have brought new challenges to the fusion task. For example, the advances in sensor technology allow better, cheaper, and smaller sensors to be used, which results in a much larger number of sensors deployed. On the other hand, sensors communicate through wireless networks where the network bandwidth is much lower than for wired communication. In this paper, we describe the usage of mobile agent for data fusion in WSNs. In this computing model, data stay at the local site, while the fusion process (code) is moved to the data sites. By transmitting the computation engine instead of data, network bandwidth requirement is largely reduced and the performance of the fusion process is more stable. One of the key problems discussed in this mobile-agentbased WSN (MAWSN) is how to plan the itinerary (or route) for a mobile agent in order to achieve progressive fusion accuracy. This paper presents a method to develop an optimal itinerary for mobile agent to fulfill the integration task while consuming minimum amount of resources, including time and power.