Efficient Heuristic for Placing Monitors on Flow Networks

摘要

Network flow monitoring is a crucial procedure for optimizing the network infrastructure for better application performance or security issue. In this paper, we address the problem of Flow Edge-Monitoring (denoted as FEM), for which we need to find k (k > 0) edges in an undirected graph G = (V, E), so that if we place k flow monitors on these edges to measure the flow along them, we can maximize the total number of edges for which the flow information can be uniquely determined according to the flow conservation property. This problem has been proved to be NP-hard, and previously reported approach suffers from higher running time complexity. In this paper, we develop efficient heuristic algorithm ALGHEU to significantly accelerating the FEM problem without compromising on solution quality, i.e. Without obviously increasing the number of required monitors. We also customize and Ant Colony algorithm, which runs slow but produce near optimal solution, to evaluate our heuristic algorithm ALGHEU. The experimental results show that, the proposed approach can significantly accelerate the previous approach by 47.67 times, while the solution quality remains nearly as good as the previous algorithm.