{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

Slides13_col

# Slides13_col - Lecture 13 The Hungarian Method The...

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

1 Lecture 13 The Hungarian Method The algorithm: Idea Given: Bipartite graph G=(V,E), edge weights c ij 0 for all (i,j) E. Find maximum weight matching in G (assignment problem). We define a (weighted) node cover : node weights y v , v Vso that y v 0 for all v Vand±y i +y j c ij for all (i,j) E. + + = M j i ij M j iv v j i V v v c y y y y ) , ( ) , ( exposed ) ( c ij 0 The main idea in the Hungarian method is to produce matchings and node coverings at the same time! Suppose M is a matching in G and y v , v V is a node cover. Then,

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

View Full Document
2 The algorithm: Idea + + = M j i ij M j iv v j i V v v c y y y y ) , ( ) , ( exposed ) ( Thus, we have Theorem: In any bipartite graph, we have Maximum matching minimum weighted node cover. Note the similarity to: 1. MaxFlow-MinCut Theorem 2. König’s Theorem If equality holds in (*), then matching is max. weight, cover is min weight. How do we get equality? (*) The algorithm: Idea + + = M j i ij M j v j i V v v c y y y y ) , ( ) , ( exposed ) ( We must have: (1) y i +y j =c ij for all (i,j) M (2) y v =0 for all exposed v V. For (1) we introduce the equality subgraph of G: G = = (V,E = ), where E = ={(i,j) E:±y i +y j =c ij } and start by finding a maximum matching in G = using the labeling algorithm.
3 The Hungarian Method • If (2) holds for this matching, then we are done!

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.

{[ snackBarMessage ]}

### Page1 / 9

Slides13_col - Lecture 13 The Hungarian Method The...

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

View Full Document
Ask a homework question - tutors are online