A multidimensional robust greedy algorithm for resource path finding in large-scale distributed networks

Abstract

This paper proposes a robust greedy algorithm and four of its variants for the resource path finding problem in distributed networks. In contrast to the existing solutions that rely on a single minimum cost path for each request, the proposed algorithm makes use of finding "robust" paths for each request within the network. We give a mathematical definition of robustness for the resource path finding problem in distributed networks. The four proposed variants are then compared with each other and with a traditional "non-robust" path finding algorithm. The simulation results show interesting improvement in solution quality when robustness is incorporated into the path finding algorithm for distributed networks.