This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: CSC 30155 Polynomial Time Reductions Dr. Ka Lok Man On Behalf of Dr. Daniel Hughes [email protected] Polynomialtime Reducibility l Another important idea in the theory of NPcompleteness is the idea of transforming one problem instance into another problem instance. We formalize this idea below. l We say that a problem L , defining some decision problem, is polynomialtime reducible to a problem M , if: there is a function f , computable in polynomial time, that takes an instance s of the problem L , and transforms it into an instance f(s) of the problem M such that s has answer “yes” if and only if f(s) has answer “yes”. l In other words, we can transform an input for one decision problem into an appropriate input for another decision problem . Furthermore, the first problem has a “yes” solution if and only if the second decision problem has a “yes” solution. l We use notation L poly M to signify that problem L is polynomial time reducible to problem M . NPhardness l We say that a problem M , defining some decision problem, is NPhard if every other problem L in NP is polynomialtime reducible to M . So this means that: l M is NPhard if for every L ∈ NP , L poly M . l If a language M is NPhard and it belongs to NP itself, then M is NPcomplete . l NPcomplete problems are some of the hardest problems in NP, as far as polynomial reducibility is concerned. Example: CookLevin Theorem l The circuit sat problem tests a given Boolean circuit for satisfiability, accepting it only if it is satisfiable. l The CookLevin Theorem states that CircuitSAT is NPcomplete. l Proof idea: The computation steps of any (reasonable) algorithm can be simulated by layers in appropriately constructed (polynomial time and size) Boolean Circuit. Other NPcomplete problems l We have just noted that there is at least one NP complete problem. l Using polynomial time reducibility we can show existence of other NPcomplete problems. l The following useful result also helps to prove NP completeness in many cases: Other NPcomplete problems l Suppose that we have a new problem X and we think that X is NPcomplete . How can we do this? l First, show that X ∈ NP , i.e. show that X has a polynomialtime nondeterministic algorithm, or equivalently, show that X has an efficient certifier. l Secondly, take a known NPcomplete problem Y , and demonstrate a polynomial time reduction from Y to X , i.e. show that Y poly X . Types of Reduction l Let M be an NPcomplete problem. There are three types of reduction: l Restriction: Noting that a known NPcomplete problem is a special case of our problem L. l Local replacement: Dividing instances of M and L into basic units, and then showing how each basic unit of M can be mapped to a basic unit of L....
View
Full
Document
This note was uploaded on 05/22/2011 for the course CSC 30155 taught by Professor Garyli during the Spring '11 term at University of Liverpool.
 Spring '11
 GaryLi
 Databases

Click to edit the document details