On optimal strategies for upgrading networks
- Wuerzburg Univ. (Germany). Dept. of Computer Science
- Los Alamos National Lab., NM (United States)
- State Univ. of New York, Albany, NY (United States). Dept. of Computer Science
- Carnegie-Mellon Univ., Pittsburgh, PA (United States). Graduate School of Industrial Administration
- Massachusetts Inst. of Tech., Cambridge, MA (United States)
We study {ital budget constrained optimal network upgrading problems}. Such problems aim at finding optimal strategies for improving a network under some cost measure subject to certain budget constraints. Given an edge weighted graph {ital G(V,E)}, in the {ital edge based upgrading model}, it is assumed that each edge {ital e} of the given network has an associated function {ital c(e)} that specifies for each edge {ital e} the amount by which the length {ital l(e)} is to be reduced. In the {ital node based upgrading model} a node {ital v} can be upgraded at an expense of cost {ital (v)}. Such an upgrade reduces the cost of each edge incident on {ital v} by a fixed factor {rho}, where 0 < {rho} < 1. For a given budget, {ital B}, the goal is to find an improvement strategy such that the total cost of reduction is a most the given budget {ital B} and the cost of a subgraph (e.g. minimum spanning tree) under the modified edge lengths is the best over all possible strategies which obey the budget constraint. Define an ({alpha},{beta})-approximation algorithm as a polynomial-time algorithm that produces a solution within {alpha} times the optimal function value, violating the budget constraint by a factor of at most {Beta}. The results obtained in this paper include the following 1. We show that in general the problem of computing optimal reduction strategy for modifying the network as above is {bold NP}-hard. 2. In the node based model, we show how to devise a near optimal strategy for improving the bottleneck spanning tree. The algorithms have a performance guarantee of (2 ln {ital n}, 1). 3. for the edge based improvement problems we present improved (in terms of performance and time) approximation algorithms. 4. We also present pseudo-polynomial time algorithms (extendible to polynomial time approximation schemes) for a number of edge/node based improvement problems when restricted to the class of treewidth-bounded graphs.
- Research Organization:
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE Assistant Secretary for Human Resources and Administration, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 435321
- Report Number(s):
- LA-UR-96-2719; CONF-970142-1; ON: DE96014461
- Resource Relation:
- Conference: 8. annual Association for Computing Machinery (ACM)-Society for Industrial and Applied Mathematics (SIAM) symposium on discrete algorithms, New Orleans, LA (United States), 5-7 Jan 1997; Other Information: PBD: 2 Jul 1996
- Country of Publication:
- United States
- Language:
- English
Similar Records
Improving spanning trees by upgrading nodes
Improving minimum cost spanning trees by upgrading nodes