Decontamination of arbitrary networks with multiple mobile agent home bases

Abstract

In this thesis we consider the problem of searching for an intruder in arbitrary networks. A team of mobile software agents is deployed to capture the intruder (i.e., a virus) and decontaminate the network. These agents follow a common protocol, derived from the Breadth-First Search (BFS) algorithm, that is independent of the intruders' speed or knowledge and is designed to prevent any further re-contamination. To measure the efficiency of our protocols and find the ones which give the best results, we conduct experiments on arbitrary synchronous and asynchronous networks. In each experiment, we study a different aspect of the impact multiple starting locations and different levels of visibility have on minimizing the number of mobile software agents required to decontaminate the network and the time used in the process.