Spanning Distribution Trees of Graphs
Let G be a graph with a single source w, assigned a positive integer called the supply. Every vertex other than w is a sink, assigned a nonnegative integer called the demand. Every edge is assigned a positive integer called the capacity. Then a spanning tree T of G is called a spanning distribution tree if the capacity constraint holds when, for every sink v, an amount of flow, equal to the demand of v, is sent from w to v along the path in T between them. The spanning distribution tree problem asks whether a given graph has a spanning distribution tree or not. In the paper, we first observe that the problem is NP-complete even for series-parallel graphs, and then give a pseudo-polynomial time algorithm to solve the problem for a given series-parallel graph G.
Keywordsspanning distribution tree series-parallel graph flow supply demand partial k-tree
Unable to display preview. Download preview PDF.
- 4.Garey, M.R., Johnson, D.S.: Computers and Intractability, A Guide to the Theory of NP-Completeness (Twenty-second printing), pp. 90–91. W.H. Freeman and Company (2000)Google Scholar
- 8.Kawabata, M., Nishizeki, T.: Partitioning trees with supply, demand and edge-capacity. In: Proc. of ISORA 2011. Lecture Notes in Operation Research, vol. 14, pp. 51–58 (2011); Also IEICE Trans. on Fundamentals of Electronics, Communications and Computer Science (to appear)Google Scholar
- 9.Kim, M.S., Lam, S.S., Lee, D.-Y.: Optimal distribution tree for internet streaming media. In: Proc. 23rd Int. Conf. on Distributed Computing System (ICDCS 2003), pp. 116–125 (2003)Google Scholar
- 10.Kleinberg, J.M.: Single-source unsplittable flow. In: Proc. of 37th FOCS, pp. 68–77 (1996)Google Scholar