skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Improved non-approximability results

Abstract

We indicate strong non-approximability factors for central problems: N{sup 1/4} for Max Clique; N{sup 1/10} for Chromatic Number; and 66/65 for Max 3SAT. Underlying the Max Clique result is a proof system in which the verifier examines only three {open_quotes}free bits{close_quotes} to attain an error of 1/2. Underlying the Chromatic Number result is a reduction from Max Clique which is more efficient than previous ones.

Authors:
;
Publication Date:
OSTI Identifier:
35820
Report Number(s):
CONF-9408161-
TRN: 94:009753-0078
Resource Type:
Conference
Resource Relation:
Conference: 15. international symposium on mathematical programming, Ann Arbor, MI (United States), 15-19 Aug 1994; Other Information: PBD: 1994; Related Information: Is Part Of Mathematical programming: State of the art 1994; Birge, J.R.; Murty, K.G. [eds.]; PB: 312 p.
Country of Publication:
United States
Language:
English
Subject:
99 MATHEMATICS, COMPUTERS, INFORMATION SCIENCE, MANAGEMENT, LAW, MISCELLANEOUS; NUMERICAL SOLUTION; EFFICIENCY; VARIATIONAL METHODS; MATRICES; LEAST SQUARE FIT

Citation Formats

Bellare, M., and Sudan, M.. Improved non-approximability results. United States: N. p., 1994. Web.
Bellare, M., & Sudan, M.. Improved non-approximability results. United States.
Bellare, M., and Sudan, M.. Sat . "Improved non-approximability results". United States. doi:.
@article{osti_35820,
title = {Improved non-approximability results},
author = {Bellare, M. and Sudan, M.},
abstractNote = {We indicate strong non-approximability factors for central problems: N{sup 1/4} for Max Clique; N{sup 1/10} for Chromatic Number; and 66/65 for Max 3SAT. Underlying the Max Clique result is a proof system in which the verifier examines only three {open_quotes}free bits{close_quotes} to attain an error of 1/2. Underlying the Chromatic Number result is a reduction from Max Clique which is more efficient than previous ones.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Sat Dec 31 00:00:00 EST 1994},
month = {Sat Dec 31 00:00:00 EST 1994}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • We study the computational complexity of the problem which consists, given a system of linear relations, of finding a solution satisfying as many relations as possible. This general combinatorial problem, referred to as maxfls, is considered for the four basic types of relations =, {ge}, > and {ne}. The problem can also be seen as a minimization problem, minulr, where we want to find a solution violating as few relations as possible. The problem is NP-hard for =, {ge} or > relations, even when restricted to homogeneous systems with bipolar coefficients, whereas it is trivial for {ne} relations. In thismore » work we determine strong bounds on the approximability of various intractable variants, including constrained ones where the variables are restricted to take bounded discrete values. We show that the approximability of different variants can differ enormously. For example maxfls{sup {ge}} and maxfls{sup >} can be approximated within a factor of 2, while maxfls{sup =}, minulr{sup =}, minulr{sup {ge}} and minulr{sup >} cannot be approximated within any constant factor. Most of the constrained variants are maxpb- or minpb-complete, which means that they are among the hardest to approximate of all optimization problems with polynomially bounded objective functions.« less
  • We investigate the complexity and approximability of some location problems when two distance values are specified for each pair of potential sites. These problems involve the selection of a specified number of facilities (i.e. a placement of a specified size) to minimize a function of one distance metric subject to a budget constraint on the other distance metric. Such problems arise in several application areas including statistical clustering, pattern recognition and load-balancing in distributed systems. We show that, in general, obtaining placements that are near-optimal with respect to the first distance metric is NP-hard even when we allow the budgetmore » constraint on the second distance metric to be violated by a constant factor. However, when both the distance metrics satisfy the triangle inequality, we present approximation algorithms that produce placements which are near-optimal with respect to the first distance metric while violating the budget constraint only by a small constant factor. We also present polynomial algorithms for these problems when the underlying graph is a tree.« less
  • The authors consider the two dimensional periodic specifications: a method to specify succinctly objects with highly regular repetitive structure. These specifications arise naturally when processing engineering designs including VLSI designs. These specifications can specify objects whose sizes are exponentially larger than the sizes of the specification themselves. Consequently solving a periodically specified problem by explicitly expanding the instance is prohibitively expensive in terms of computational resources. This leads one to investigate the complexity and efficient approximability of solving graph theoretic and combinatorial problems when instances are specified using two dimensional periodic specifications. They prove the following results: (1) several classicalmore » NP-hard optimization problems become NEXPTIME-hard, when instances are specified using two dimensional periodic specifications; (2) in contrast, several of these NEXPTIME-hard problems have polynomial time approximation algorithms with guaranteed worst case performance.« less
  • We consider the problem of fitting an n {times} n distance matrix D by a tree metric T. Let {epsilon} be the distance to the closest tree metric under the L{sub {infinity}} norm, that is, {epsilon} = min{sub T} ({parallel}T, D{parallel}{sub {infinity}}). First we present an O(n{sup 2}) algorithm for finding an additive tree T such that {parallel}T, D {parallel}{sub {infinity}} {le} 3{epsilon}. Second we show that it is NP-hard to find a tree T such that {parallel} T, D {parallel} {sub {infinity}} < 9/8{epsilon}. This paper presents the first algorithm for this problem with a performance guarantee.