# Lec4 - COT 6936: Topics in Algorithms Giri Narasimhan ECS...

This preview shows pages 1–4. Sign up to view the full content.

1 1/12/10 COT 6936 1 COT 6936: Topics in Algorithms Giri Narasimhan ECS 254A / EC 2443; Phone: x3748 [email protected] http://www.cs.fiu.edu/~giri/teach/COT6936_S10.html https://online.cis.fiu.edu/portal/course/view.php?id=427 Optimization Problems • Problem: – A problem is a function (relation) from a set I of instances of the problem to a set S of solutions. p: I ± S • Decision Problem: – Problem with S = {TRUE, FALSE} • Optimization Problem: – Problem with a mapping from set S of solutions to a positive rational number called the solution value I ± S ± m(I,S) 1/12/10 COT 6936 2 Optimization Versions of NP-Complete Problems • TSP • CLIQUE • Vertex Cover & Set Cover • Hamiltonian Cycle • Hamiltonian Path • SAT & 3SAT • 3-D matching 1/12/10 COT 6936 3

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
2 Optimization Versions of NP-Complete Problems • Computing a minimum TSP tour is NP-hard (every problem in NP can be reduced to it in polynomial time) • BUT, it is not known to be in NP • If P is NP-Complete, then its optimization version is NP-hard (i.e., it is at least as hard as any problem in NP, but may not be in NP) – Proof by contradiction! 1/12/10 COT 6936 4 Performance Ratio • Approximation Algorithm A – A( I ) • Optimal Solution – OPT( I ) • Performance Ratio on input I for minimization problems – R A ( I ) = max {A( I )/OPT( I ), OPT( I )/A( I )} • Performance Ratio of approximation algorithm A – R A = inf {r ± 1| R A ( I ) ² r, for all instances} 1/12/10 COT 6936 5 Metric Space • It generalizes concept of Euclidean space • Set with a distance function (metric) defined on its elements – D : M X M R (assigns a real number to distance between every pair of elements from the metric space M ) • D( x , y ) = 0 iff x = y • D( x , y ) ± 0 • D( x , y ) = D( y , x ) • D( x , y ) + D( y , z ) ± D( x , z ) 1/12/10 COT 6936 6
Examples of metric spaces • Euclidean distance • L p metrics • Graph distances – Distance between elements is the length of the shortest path in the graph 1/12/10 COT 6936 7 TSP • TSP in general graphs cannot be approximated to within a constant ( Why ?) – What is the approach? • Prove that it is hard to approximate! • TSP in general metric spaces holds promise! – NN heuristic [Rosenkrantz, et al. 77] • NN(I) ± ² (ceil(log 2 n) + 1) OPT(I) – 2-OPT, 3-OPT, k-OPT, Lin-Kernighan Heuristic • Can TSP in general metric spaces be approximated to within a constant? 1/12/10 COT 6936 8 TSP in Euclidean Space • TSP in Euclidean space can be approximated. – MST Doubling (DMST) Algorithm • Compute a MST, M • Double the MST to create a tour, T 1 • Modify the tour to get a TSP tour, T – Theorem : DMST is a 2-approximation algorithm for Euclidean metrics, i.e., DMST( I ) < 2 OPT( I ) – Analysis: • L(T) ± L(T 1 ) = 2L(M) ± 2L(T OPT ) – Is the analysis tight? 1/12/10

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 02/18/2012 for the course CIS 6936 taught by Professor Giri during the Spring '12 term at FIU.

### Page1 / 16

Lec4 - COT 6936: Topics in Algorithms Giri Narasimhan ECS...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online