Editing Open Problems:60
Warning: You are not logged in. Your IP address will be publicly visible if you make any edits. If you log in or create an account, your edits will be attributed to your username, along with other benefits.
The edit can be undone.
Please check the comparison below to verify that this is what you want to do, and then save the changes below to finish undoing the edit.
Latest revision | Your text | ||
Line 1: | Line 1: | ||
{{Header | {{Header | ||
+ | |title=Single-Pass Unweighted Matchings | ||
|source=dortmund12 | |source=dortmund12 | ||
|who=Andrew McGregor | |who=Andrew McGregor | ||
}} | }} | ||
Suppose you have $O(n \operatorname{polylog} n)$ memory and a single pass over a stream of $m$ edges (arbitrarily ordered) on $n$ nodes. How well can you approximate the size of the maximum cardinality matching? A trivial greedy algorithm finds a $1/2$-approximation but that's still the best known algorithm in the general setting. Kapralov {{cite|Kapralov-12}} showed that achieving better than a $1-1/e$ approximation is impossible. If the stream is randomly ordered, Konrad et al. {{cite|KonradMM-12}} presented a $1/2 + 0.005$-approximation. Other variants of the question are also open, e.g., achieving a $(1-\epsilon)$ approximation in the minimum number of passes (see, e.g., Ahn and Guha {{cite|AhnG-11}}) or the best approximation possible for maximum weighted matching in a single pass (see, e.g., Epstein et al. {{cite|EpsteinLMS-11}}). | Suppose you have $O(n \operatorname{polylog} n)$ memory and a single pass over a stream of $m$ edges (arbitrarily ordered) on $n$ nodes. How well can you approximate the size of the maximum cardinality matching? A trivial greedy algorithm finds a $1/2$-approximation but that's still the best known algorithm in the general setting. Kapralov {{cite|Kapralov-12}} showed that achieving better than a $1-1/e$ approximation is impossible. If the stream is randomly ordered, Konrad et al. {{cite|KonradMM-12}} presented a $1/2 + 0.005$-approximation. Other variants of the question are also open, e.g., achieving a $(1-\epsilon)$ approximation in the minimum number of passes (see, e.g., Ahn and Guha {{cite|AhnG-11}}) or the best approximation possible for maximum weighted matching in a single pass (see, e.g., Epstein et al. {{cite|EpsteinLMS-11}}). |