Cost Effective Mobile Agent Planning for Distributed Information Retrieval
ICDCS '01 Proceedings of the The 21st International Conference on Distributed Computing Systems
Ant system: optimization by a colony of cooperating agents
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
| Hi-index | 0.00 |
Mobile agents often have the task to collect data from several predefined sites. This should be done in an efficient way by minimizing the elapsed time. Usually the agents only know the list of sites but not the distances between them. In contrast to former publications this paper presents an algorithm to determine the agent's tour which does not depend on a known distance matrix. The agent repeats traveling over the short routes and avoids longer ones. The paper shows that the optimal threshold point depends on the number of visits the agent takes or, equivalently, the number of tours taken within that period where the collected network information does not get outdated. Using the proposed algorithm the total latency time can be halved (against the common strategy of a random way) when taking 30 runs, without producing additional network traffic. These results are derived analytically and verified by simulation.