Improving Steiner trees of a network under multiple constraints
- Univ. of Wuerzburg (Germany). Dept. of Computer Science
- Los Alamos National Lab., NM (United States)
- Carnegie Mellon Univ., Pittsburgh, PA (United States)
- State Univ. of New York, Albany, NY (United States). Dept. of Computer Science
The authors consider the problem of decreasing the edge weights of a given network so that the modified network has a Steiner tree in which two performance measures are simultaneously optimized. They formulate these problems, referred to as bicriteria network improvement problems, by specifying a budget on the total modification cost, a constraint on one of the performance measures and using the other performance measure as a minimization objective. Network improvement problems are known to be NP-hard even when only one performance measure is considered. The authors present the first polynomial time approximation algorithms for bicriteria network improvement problems. The approximation algorithms are for two pairs of performance measures, namely (diameter, total cost) and (degree, total cost). These algorithms produce solutions which are within a logarithmic factor of the optimum value of the minimization objective while violating the constraints only by a logarithmic factor. The techniques also yield approximation schemes when the given network has bounded treewidth. Many of the approximation results can be extended to more general network design problems.
- Research Organization:
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 251414
- Report Number(s):
- LA-UR-96-1374; CONF-960882-1; ON: DE96010936; TRN: AHC29614%%142
- Resource Relation:
- Conference: 4. Europeoan symposium on algorithms, Corfu (Greece), Aug 1996; Other Information: PBD: [1996]
- Country of Publication:
- United States
- Language:
- English
Similar Records
Bicriteria network design problems
Bicriteria network design problems