AIAug 27, 2012

Distributed Pharaoh System for Network Routing

arXiv:1208.5373v11 citations
Originality Synthesis-oriented
AI Analysis

This work addresses network routing optimization for communication systems, but it is incremental as it builds upon existing AntNet and Pharaoh Ant System methods.

The paper tackles the Low Cost Network Routing problem by introducing the Distributed Pharaoh System (DPS), a biobjective ant algorithm that converges to the shortest path and low-cost overlay routing network topology, with numerical experiments on a random 10-node network showing encouraging results.

In this paper it is introduced a biobjective ant algorithm for constructing low cost routing networks. The new algorithm is called the Distributed Pharaoh System (DPS). DPS is based on AntNet algorithm. The algorithm is using Pharaoh Ant System (PAS) with an extra-exploration phase and a 'no-entry' condition in order to improve the solutions for the Low Cost Network Routing problem. Additionally it is used a cost model for overlay network construction that includes network traffic demands. The Pharaoh ants (Monomorium pharaonis) includes negative pheromones with signals concentrated at decision points where trails fork. The negative pheromones may complement positive pheromone or could help ants to escape from an unnecessarily long route to food that is being reinforced by attractive signals. Numerical experiments were made for a random 10-node network. The average node degree of the network tested was 4.0. The results are encouraging. The algorithm converges to the shortest path while converging on a low cost overlay routing network topology.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes