A diffusional distributed multiple path algorithm for complex networks

This paper proposes a new algorithm to compute the distributed multiple path (DMP) sets in a network, where the DMP set of an O-D (origin-destination) pair in the network contains (if possible) multiple paths from the origin node to the destination node, which all possess short length and meanwhile overlap relatively less with each other. As during the computation, the algorithm relies on a key operation that, each node broadcasts (or to say, diffuses) its newly computed DMPs to its neighbor nodes, and thus the algorithm is named as the diffusional distributed multiple path algorithm (DDMPA). We test DDMPA via computer simulation based on the BA scale-free network model. The results show that, DDMPA can calculate the DMP sets that can better support the distributed multiple path routing method, and meanwhile it has a low time complexity that makes it possible to be applied in large-scale networks.