Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
11 views

Combinatorial Optimization: Sheet 5

This document contains 6 exercises on the topic of combinatorial optimization. It provides instructions for students to receive bonus credit by submitting written solutions to problems marked with an asterisk. The exercises cover topics like independence systems, forests in graphs, matroids, submodular functions, and computing minimum weight arborescences.

Uploaded by

Anonymous ewKJI4xU
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
11 views

Combinatorial Optimization: Sheet 5

This document contains 6 exercises on the topic of combinatorial optimization. It provides instructions for students to receive bonus credit by submitting written solutions to problems marked with an asterisk. The exercises cover topics like independence systems, forests in graphs, matroids, submodular functions, and computing minimum weight arborescences.

Uploaded by

Anonymous ewKJI4xU
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Prof. F.

Eisenbrand EPFL - DISOPT

Combinatorial Optimization

Adrian Bock Fall 2011

Sheet 5 November 17, 2011

General remark:
In order to obtain a bonus for the final grading, you may hand in written solutions
to the exercises marked with a star at the beginning of the exercise session on
November 29.

Exercise 1

Let M = (E, I) be an independence system (i.e. (M0) and (M1) hold).


Prove or give a counterexample that the following two properties are equivalent:
(M2) For all I, J ∈ I and |J| > |I| there exists some e ∈ J \I such that I∪{e} ∈ I.
(M2′ ) For all I, J ∈ I with I ∩ J = ∅ and |J| > |I|, there exists some e ∈ J \ I
such that I ∪ {e} ∈ I.

Exercise 2

Let G = (V, E) be a graph. Show that conv({χF ∈ {0, 1}|E| : F is a forest in G}),
the convex hull of incidence vectors of all forests in G is equal to the following set
|E|
X
{x ∈ R≥0 : x(e) ≤ |T | − 1, for all ∅ =6 T ⊆ V}
e∈E[T ]

Exercise 3 (⋆)

Let E be a finite set and let r : 2E → Z+ . Then r is the rank function of a matroid
(E, I) if and only if for all I, J ⊆ E :
(i) r(I) ≤ r(J) ≤ |J| if I ⊆ J ,
(ii) r(I ∩ J) + r(I ∪ J) ≤ r(I) + r(J).
Hint: "⇐" Show first that (E, I) where I = {I ⊆ E : |I| = r(I)} is a matroid
(using condition (iii) in exercise 2 on sheet 4). Conclude that r is indeed a rank
function by induction on the size of A ⊆ E.
Exercise 4

We consider the following generalization of the matroid polytope:


Let E be a finite set and let f : 2E → R+ be a submodular function, i.e.
f (I ∩ J) + f (I ∪ J) ≤ f (I) + f (J) for all I, J ⊆ E. The polymatroid defined by
f is the polytope
( )
X
P (f ) := x ∈ RE : x ≥ 0, xe ≤ f (A) ∀A ⊆ E
e∈A

Note that the matroid polytope is a special case where f is the rank function of
a matroid over E. What is the separation problem for a polymatroid?
Let G = (V, E) be aPgraph and c : E → R+ . Consider the function
f (A) := c(δ(A)) = e∈δ(A) c(e) for all A ⊆ V .
(i) Show that f is submodular.
(ii) Show how to solve the separation problem for the polymatroid defined by
f over V in time O(m2 n2 ).
(iii) Conclude how to maximize a weight function w : V → R over the polyma-
troid P (f ).

Exercise 5
Trace the steps of algorithm from the lecture to compute a minimum weight
arborescence rooted at r in the following example.
r
5 2

1
a b

4
3 5 2

c d
4

Exercise 6 (⋆)
Let D = (V, A) be a directed graph with root r ∈ V . Suppose that D does
not contain an arborescence rooted at r. Prove that there exists a (nonempty)
strongly connected component K in D such that r ∈/ K and |δ in (K)| = 0.

You might also like