RUS  ENG
Full version
JOURNALS // Zapiski Nauchnykh Seminarov POMI // Archive

Zap. Nauchn. Sem. POMI, 2015 Volume 436, Pages 122–135 (Mi znsl6163)

This article is cited in 2 papers

On a class of optimization problems with no “effectively computable” solution

M. R. Gavrilovich, V. L. Kreps

National Research University "Higher School of Economics", St. Petersburg Branch

Abstract: It is well known that large random structures may have nonrandom macroscopic properties. We give an example of nonrandom properties for a class of large optimization problems related to the computational problem $MAX\,FLS^=$ of calculating the maximum number of consistent equations in a given overdetermined system of linear equations.
For this class we establish the following. There is no “efficiently computable” optimal strategy. When the size of a random instance of the optimization problem goes to infinity, the probability that the uniform mixed strategy is $\varepsilon$-optimal goes to one. Moreover, there is no “efficiently computable” strategy that is substantially better for each instance of the optimization problem.

Key words and phrases: optimization, concentration of measure, probabilistically checkable proofs.

UDC: 519.248.6+519.87+519.876+519.83

Received: 02.09.2015


 English version:
Journal of Mathematical Sciences (New York), 2016, 215:6, 706–714

Bibliographic databases:


© Steklov Math. Inst. of RAS, 2024