Editing Open Problems:71

Jump to: navigation, search

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=Metric TSP Cost Approximation
 
|source=baltimore16
 
|source=baltimore16
 
|who=Krzysztof Onak
 
|who=Krzysztof Onak
 
}}
 
}}
This problem appeared in the paper of Czumaj and Sohler {{cite|CzumajS-09}}, who gave an efficient algorithm for approximating the weight of the minimum spanning tree of a metric space in the following setting. The input is a metric space $\mathcal M$ on $n$ points. For any pair of points in $\mathcal M$, the algorithm is allowed to query their distance. The algorithm of Czumaj and Sohler computes a multiplicative $(1+\epsilon)$-approximation to the ''weight'' of the minimum spanning tree in time $\tilde O(n)\cdot\operatorname{poly}(1/\epsilon)$. This immediately yields a $(2+\epsilon)$-approximation to the length of the minimum travelling salesman tour.
+
This problem appeared in the paper of Czumaj and Sohler {{cite|CzumajS-09}}, who gave an efficient algorithm for approximating the weight of the minimum spanning tree of a metric space in the following setting. The input is a metric space $\mathcal M$ on $n$ points. For any pair $u$ and $v$ of points in $\mathcal M$, the algorithm is allowed to query their distance. The algorithm of Czumaj and Sohler computes a multiplicative $(1+\epsilon)$-approximation to the weight of the minimum spanning tree in time $\tilde O(n)\cdot\operatorname{poly}(1/\epsilon)$. This immediately yields a $(2+\epsilon)$-approximation to the length of the minimum travelling salesman tour.
  
'''Question:''' Can one compute a $(2-\delta)$-approximation to the length of the minimum travelling salesman tour with $o(n^2)$ queries or—even better—in $o(n^2)$ time, for a positive absolute constant $\delta$?
+
'''Question:''' Can one compute a $(2-\delta)$-approximation to this problem with $o(n^2)$ queries (or even better, in $o(n^2)$ time) for a positive absolute constant $\delta$?
  
 
'''Notes:'''
 
'''Notes:'''
* It is difficult to apply the approach known from the Christofides algorithm, which computes a $3/2$-approximation in the more standard setting. It requires computing a minimum spanning tree, while Czumaj and Sohler show that obtaining a constant-factor approximation to the minimum spanning tree (''not just its weight'') requires $\Omega(n^2)$ queries.
+
* It is difficult to apply the approach known from the Christofides algorithm, which computes a $3/2$-approximation in the more standard setting. It requires computing a minimum spanning tree, while Czumaj and Sohler show that obtaining a constant-factor approximation to the minimum spanning tree requires $\Omega(n^2)$ queries.
  
* For the specific case of $(1,2)$-metrics (i.e., when all distances between points are either 1 or 2), there is an $\tilde O(n)\cdot\operatorname{poly}(1/\epsilon)$-time $(1.75+\epsilon)$-approximation algorithm. It uses the fact that if there is a short travelling salesman tour, then there must be a large matching in the graph obtained by connecting pairs of points at distance 1. Therefore, the cases of long and short shortest tours can be distinguished efficiently using a known graph matching algorithm {{cite|OnakRRR-12}}. (This is a joint observation with Anupam Gupta.) However, this approach does not seem to generalize to arbitrary metrics.
+
* For the specific case of $(1,2)$-metrics (i.e., when all distances between points are either 1 or 2), there is a $\tilde O(n)\cdot\operatorname{poly}(1/\epsilon)$-time $(1.75+\epsilon)$-approximation algorithm. It uses the fact that if there is a short travelling salesman tour, then there must be a large matching in the graph obtained by connecting pairs of points at distance 1. Therefore, the cases of long and short shortest tours can be distinguished efficiently using a known graph matching algorithm {{cite|OnakRRR-12}}. (This is a joint observation with Anupam Gupta.) However, this approach does not seem to generalize to arbitrary metrics.

Please note that all contributions to Open Problems in Sublinear Algorithms may be edited, altered, or removed by other contributors. If you do not want your writing to be edited mercilessly, then do not submit it here.
You are also promising us that you wrote this yourself, or copied it from a public domain or similar free resource (see Open Problems in Sublinear Algorithms:Copyrights for details). Do not submit copyrighted work without permission!

To edit this page, please answer the question that appears below (more info):

Cancel Editing help (opens in new window)