dc.description.abstract | The (n, k)-arrangement graph was first introduced in 1992 as a generalization of the star graph topology. Choosing an arrangement topology is more efficient in comparison with a star graph as we can have a closer number of nodes to what is needed. Also it has other advantages such as a lower degree and a smaller diameter, depending on k. In this thesis we investigate the problem of finding k(n − k) disjoint paths from a source node to k(n−k) target nodes in an (n, k)-arrangement interconnection network such that no path has length more than diameter+(n−k)+2, where diameter is the maximum length of shortest path between any two nodes in the graph. These disjoint paths are built by routing to all neighbors of the source node and fixing specific elements in each of the k positions of the node representation in an (n, k)-arrangement graph. Moreover, a simple routing is presented for finding n disjoint paths between two nodes which are located in different sub-graphs. The lengths are no more than d(t, s) + 4, for d(t, s) being the shortest path length between two nodes s and t. This routing algorithm needs O(n^2) time to find all n these paths.
In addition to arrangement graphs, we also study augmented cubes, first introduced in 2002, a desirable variation of the hypercube. An augmented cube of dimension n has a higher degree and a lower diameter in comparison with the hypercube. We introduce an O(n^3) algorithm for finding disjoint shortest paths from a single source node to 2n − 1 different target nodes. | en_US |