Linear Programming and Maximum Flow Problem, Lecture notes of Convex Optimization

A series of homework problems related to linear programming (lp) and the maximum flow problem. The first problem asks to formulate the problem of finding the best fitting line through a set of data points as an lp and discusses the reason for solving the dual. The second problem involves maximizing a linear function subject to certain constraints using the simplex method. The third problem asks if a pivot can move the feasible point without changing the cost and if a vector can re-enter the basis on the next pivot. The fourth problem deals with the dual of the maximum flow problem and shows how to construct a cut of minimum capacity separating the source and target nodes using the optimal solution.

Typology: Lecture notes

2012/2013

Uploaded on 04/23/2013

ashwini
ashwini 🇮🇳

4.5

(18)

167 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
HW 4: due Tues, October 17th
1. A physicist takes measurements of a variable y(x), the results are in
the form of pairs (xi, yi). The physicist wishes to find the straight
line that fits this data best, in the sense that the maximum vertical
distance between any point (xi, yi) and the line is as small as possible.
Formulate this problem as an LP. Why might you decide to solve the
dual?
2. Solve the following problem by the simplex method:
Maximize 3x1+ 2x2+ 4x3
subject to
x1+x2+ 2x34
2x1+ 3x35
2x1+x2+ 3x37
x1, x2, x30
3. Answer yes or no and prove your answer: can a pivot of the simplex
algorithm move the feasible point a positive distance in Rnwhile leaving
the cost unchanged?
4. Can a vector which has just left the basis in the simplex algorithm
re-enter on the very next pivot?
pf2

Partial preview of the text

Download Linear Programming and Maximum Flow Problem and more Lecture notes Convex Optimization in PDF only on Docsity!

HW 4: due Tues, October 17th

  1. A physicist takes measurements of a variable y(x), the results are in the form of pairs (xi, yi). The physicist wishes to find the straight line that fits this data best, in the sense that the maximum vertical distance between any point (xi, yi) and the line is as small as possible. Formulate this problem as an LP. Why might you decide to solve the dual?
  2. Solve the following problem by the simplex method:

Maximize 3x 1 + 2x 2 + 4x 3

subject to

x 1 + x 2 + 2x 3 ≤ 4

2 x 1 + 3x 3 ≤ 5 2 x 1 + x 2 + 3x 3 ≤ 7 x 1 , x 2 , x 3 ≥ 0

  1. Answer yes or no and prove your answer: can a pivot of the simplex algorithm move the feasible point a positive distance in Rn^ while leaving the cost unchanged?
  2. Can a vector which has just left the basis in the simplex algorithm re-enter on the very next pivot?
  1. The maximum flow problem on the directed graph G = (V, E) with capacity function u (and lower bounds 0) can be formulated by the following linear program: maximize w subject to ∑ j

xij −

∑ j

xji = w if i = s ; 0, if i 6 = s, t; −w if i = t

xij ≤ uij

0 ≤ xij

(a) Show that its dual is equivalent to:

min

∑ (i,j)∈E

uij yij

subject to

zi − zj + yij ≥ 0 (i, j) ∈ E

zs = 0, zt = 1

yij ≥ 0

(b) (extra credit-hard) Given any (not necessarily integral) optimal solution y∗, z∗^ of value W ∗^ to this dual linear program, show how to construct from z∗^ a cut separating s from t of value W ∗. Hint: make a clever use of complementary slackness (c) Deduce the max-flow–min-cut theorem: the value of the maximum flow from s to t is equal to th evalue of the minimum cut separating s and t.