Fault-Tolerant Version - Advanced Algorithms - Exam, Exams of Advanced Algorithms

Main points of this exam paper are: Fault-Tolerant Version, K-Center Problem, Triangle Inequality, Additional Input, Fault-Tolerant Cost, Dominating Set, Undirected Graph, Size of Minimum Cardinality, Approximation Algorithm, Perfect Matching

Typology: Exams

2012/2013

Uploaded on 04/23/2013

atasi
atasi 🇮🇳

4.6

(32)

134 documents

1 / 1

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
HW 4: due Thurs, March 10
NOTE: No class on Thurs, Feb 24: Tufts is on a MONDAY schedule
Also, do problem 2 first; you won’t be able to do Problem 1 until
after the March 3 lecture.
1. The fault-tolerant version of the k-center problem with triangle inequal-
ity has an additional input αkwhich specifies the number of centers
that each vertex must be connected to. In other words, we assume that
up to α1 centers might be closed, and so the fault-tolerant cost for
a vertex is its distance to its αth closest center. The problem is to
pick kcenters so that the maximum fault-tolerant cost of a vertex is
minimized. A set SVin an undirected graph H= (V, E) is an α-
dominating set if each vertex vVis adjacent to at least αvertices in
S(we consider a vertex to be adjacent to itself). Let domα(H) denote
the size of a minimum cardinality α-dominating set in H.
(a) Let Ibe an independent set in H2. Show that α|I| domα(H).
(b) Give a factor 3 approximation algorithm for the fault-tolerant k-
center problem (Hint: Compute a maximal independent set Miin
G2
i, for 1 im. Find the smallest index isuch that |Mi| b k
αc,
and moreover, the degree of each vertex of Miin Giis α1.)
2. Let Gbe a complete undirected graph in which all edge lengths are
either 1 or 2.
(a) Show Gsatisfies the triangle inequality.
(b) () Show that you can find a minimum 2-factor in polynomial
time in this graph, where a 2-factor is a subgraph in which every
vertex is incident to exactly two edges in the 2-factor. (Remark:
the definition of k-factor is obvious; a perfect matching can also
be called a 1-factor).
(c) Give a polynomial-time algorithm to find a 4/3-approximation to
the optimum TSP in G. (You may use part b even if you were
not able to do it).

Partial preview of the text

Download Fault-Tolerant Version - Advanced Algorithms - Exam and more Exams Advanced Algorithms in PDF only on Docsity!

HW 4: due Thurs, March 10

NOTE: No class on Thurs, Feb 24: Tufts is on a MONDAY schedule Also, do problem 2 first; you won’t be able to do Problem 1 until after the March 3 lecture.

  1. The fault-tolerant version of the k-center problem with triangle inequal- ity has an additional input α ≤ k which specifies the number of centers that each vertex must be connected to. In other words, we assume that up to α − 1 centers might be closed, and so the fault-tolerant cost for a vertex is its distance to its αth closest center. The problem is to pick k centers so that the maximum fault-tolerant cost of a vertex is minimized. A set S ⊆ V in an undirected graph H = (V, E) is an α- dominating set if each vertex v ∈ V is adjacent to at least α vertices in S (we consider a vertex to be adjacent to itself). Let domα(H) denote the size of a minimum cardinality α-dominating set in H.

(a) Let I be an independent set in H^2. Show that α|I| ≤ domα(H). (b) Give a factor 3 approximation algorithm for the fault-tolerant k- center problem (Hint: Compute a maximal independent set Mi in G^2 i , for 1 ≤ i ≤ m. Find the smallest index i such that |Mi| ≤ b (^) αk c, and moreover, the degree of each vertex of Mi in Gi is ≥ α − 1.)

  1. Let G be a complete undirected graph in which all edge lengths are either 1 or 2.

(a) Show G satisfies the triangle inequality. (b) (∗) Show that you can find a minimum 2-factor in polynomial time in this graph, where a 2-factor is a subgraph in which every vertex is incident to exactly two edges in the 2-factor. (Remark: the definition of k-factor is obvious; a perfect matching can also be called a 1-factor). (c) Give a polynomial-time algorithm to find a 4/3-approximation to the optimum TSP in G. (You may use part b even if you were not able to do it).