Accelerated Diffusion Algorithms on General Dynamic Networks

In this paper, an adaptation of the first order diffusive scheme to dynamic networks and its accelerated version with an optimal relaxation parameter are given. A network with dynamic links assumes that edges of the network topology may be unpredictably broken down. Contrary to the static network framework, the connection graphs of the dynamic network may be always not connected. Various simulations on static and dynamic networks confirm our theoretical results.

[1]  Jacques M. Bahi,et al.  Load Balancing on Networks with Dynamically Changing Topology , 2001, Euro-Par.

[2]  Bruce M. Maggs,et al.  On the Fault Tolerance of Some Popular Bounded-Degree Networks , 1998, SIAM J. Comput..

[3]  Robert J. Plemmons,et al.  Nonnegative Matrices in the Mathematical Sciences , 1979, Classics in Applied Mathematics.

[4]  George Cybenko,et al.  Dynamic Load Balancing for Distributed Memory Multiprocessors , 1989, J. Parallel Distributed Comput..

[5]  Jacques M. Bahi,et al.  Broken edges and dimension exchange algorithms on hypercube topology , 2003, Eleventh Euromicro Conference on Parallel, Distributed and Network-Based Processing, 2003. Proceedings..

[6]  Seyed Hossein Hosseini,et al.  Analysis of a Graph Coloring Based Distributed Load Balancing Algorithm , 1990, J. Parallel Distributed Comput..