Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Quantum complexity theory
Bernstein E., Vazirani U. SIAM Journal on Computing26 (5):1411-1473,1997.Type:Article
Date Reviewed: Mar 1 1999

The authors define a quantum Turing machine (QTM) analogously to the classic Turing machine as a triplet ( &Sgr; , Q , &dgr; ), where &Sgr; is a finite alphabet with an identified blank symbol #, Q is a finite set of states with an identified initial state q0 and a final state qf ≠ q0, and &dgr;, the quantum transition function, is a function from Q × &Sgr; into the set of functions from &Sgr; × Q × { L, R } to &Ctilde;, where &Ctilde; is a subset of complex numbers whose real and imaginary parts can be deterministically calculated to within 2-n in time polynomial in n. The QTM operates on a two-way infinite tape of cells with a single read/write head. Configurations, initial configurations, and final configurations are defined as for classic deterministic Turing machines. A superposition is an element of a set of finite complex linear combinations of configurations, which forms an inner-product space using the Euclidean norm. The QTM defines a linear operator, called the time evolution operator, on the set of superpositions S. A QTM is well-formed if and only if its time evolution operator is unitary. A QTM is in normal form if the machine steps right and leaves the tape unchanged in moving from state qf to q0. Well-behaved normal form QTMs can be dovetailed to form a normal-form QTM that carries out the computation of the first followed by the computation of the second. The authors provide a construction of a universal QTM in Theorem 7.1. “There is a normal form QTM M such that for any well-formed QTM, any &egr; > 0, and any T, M can simulate the QTM with accuracy &egr; for T steps with slowdown polynomial in T and 1 &slash; &egr;.”

The remainder of the paper defines languages accepted by QTMs and classes of languages accepted by some deterministic polynomial-time QTM. In particular, the class P of languages accepted by polynomial-time Turing machines is in the class EQP of languages accepted by error-free quantum polynomial time QTMs, and the class BPP is contained in the class BQP of languages that are accepted with probability by some polynomial-time QTM.

Reviewer:  J. S. Williams Review #: CR121354 (9903-0184)
Bookmark and Share
 
Bounded-Action Devices (F.1.1 ... )
 
 
General (G.2.0 )
 
 
Physical Sciences And Engineering (J.2 )
 
Would you recommend this review?
yes
no
Other reviews under "Bounded-Action Devices": Date
Fairness and conspiracies
Best E. Information Processing Letters 18(4): 215-220, 1984. Type: Article
May 1 1985
Separating the eraser Turing machine classes Le, NLe, co-NLe and Pe
Krause M., Meinel C., Waack S. Theoretical Computer Science 86(2): 267-275, 1991. Type: Article
Jul 1 1992
Two nonlinear lower bounds for on-line computations
Dūris P., Galil Z., Paul W. (ed), Reischuk R. Information and Control 60(1-3): 1-11, 1984. Type: Article
Aug 1 1985
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy