Scribd
Upload
40 views3 pages

Preliminary Solutions To Examination TNK053 2009-12-21 1. 2

This document summarizes the solutions to several optimization problems: 1) Finds the optimal solution to a linear program with variables y1, y2, y3 bounded between given values. 2) Solves the dual of another linear program to find the optimal dual solution. 3) Shows that a new constraint added to the original problem does not change the optimal primal solution. It then summarizes the solutions to several other optimization problems involving finding shortest paths in graphs, solving a linear program to minimize transportation costs, using Gomory cuts in integer programming, applying KKT conditions to find the minimum of a convex problem, and solving a binary linear program using its dual.

Uploaded by

Fernanda Paula Rocha
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
40 views3 pages

Preliminary Solutions To Examination TNK053 2009-12-21 1. 2

This document summarizes the solutions to several optimization problems: 1) Finds the optimal solution to a linear program with variables y1, y2, y3 bounded between given values. 2) Solves the dual of another linear program to find the optimal dual solution. 3) Shows that a new constraint added to the original problem does not change the optimal primal solution. It then summarizes the solutions to several other optimization problems involving finding shortest paths in graphs, solving a linear program to minimize transportation costs, using Gomory cuts in integer programming, applying KKT conditions to find the minimum of a convex problem, and solving a binary linear program using its dual.

Uploaded by

Fernanda Paula Rocha
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Preliminary Solutions to Examination TNK053 2009-12-21

1.

(i) (2p) y1 =

2.

(i) (2p)

11
14 ,

y2 =

1
35 ,

y3 = 0

max
s.t.

(ii) (2p) 128 b2 240

(iii) (2p) c3

46
7

w = 10y1 + 2y2
y1 + 2y2 + y3 3
2y1 y3 1
y1 y2 2
y1 0, y2 urs , y3 0

(ii) (2p) By the complementary slackness


and w = z = 2

Optimal dual solution: (y1 , y2 , y3 ) = (0, 1, 1),

(iii) (2p) The current optimal solution solution (x1 , x2 , x3 ) = (1, 1, 0) still satisfies the new
constraint 4x1 x2 + x3 5, so the optimal primal solution is not changed. Since the slack
variable for the new constraint is not equal to zero, then the dual variable y4 = 0. the
optimal dual solution: (y1 , y2 , y3 , y4 ) = (0, 1, 1, 0).

3.

(i) (3p) The shortest path is 1-2-4-6 with cost -5.

(ii) (3p) Send 2 units from node 1 to node 6 through the shortest path 1-2-4-6, and then send
2 units through the new shortest path 1-3-4-6. The total cost is -16.( 2(-5)+ 2(-3)= -16).

4.

(6p)

Define
xij = the no. of barrels of oil shipped from location i to j per day. Assume that all flows xij are
measured in units of 100,000 barrels.
min

300x13 + 110x14 + 420x23 + 100x24 + (700 + 450)x35


+(700 + 550)x36 + (900 + 470)x45 + (900 + 530)x46

s.t.

x13 x14 = 5
x23 x24 = 4
x13 + x23 x35 x36 x37 = 0
x14 + x24 x45 x46 x47 = 0
x35 + x45 = 4
x36 + x46 = 3
x37 + x47 = 2
xij 0

i, j

D
F1
500,000

C1
400,000

F2
400,000

C2
300,000

H
7

Dummy
200,000

5.
(a).

(i) (2p) The Gomorys fractional cut associated with x1 : 13 x3 32 x4 23 x2 6


The Gomorys fractional cut associated with x5 : 23 x3 13 x4 13 x1 +x2 8
(ii) (1p) No. it cut away the feasible integer solution, for example (0,4).

(b). (3p)
~(0)
x = (0, 0.5, 1, 0)
~(0)
z = 11
x~(1)= (0, 1, 0.4, 0)
~(1)
z = 14

x 2 =1

x2 = 0

x3 = 1
2
F
~ (2)
x = (0, 1, 1, 0)
~ (2)
^
z = 17= z

x3 = 0

~(4)
x = (1, 0, 1, 0)
~
z (4) = 13 < 17

OPT.

3
F
~(3)
x = (2/3, 1, 0, 0)
~
z (3) = 17.33 > 17
by bound

6.

(i) (2p) Since f (x) is convex for x 0, and gi (x), i = 1, 2, 3, 4 are convex, so the problem
is a convex problem.
(ii) (2p) The KKT conditions:
(1)

f (x) +

4
X

ui gi (x) 0

(2)

xj

i=1

(3)

gi (x) bi ,

(4)

ui (bi gi (x)) = 0,

(5)

xj 0,

(6)

ui 0,

j = 1, 2

where f = (3(x1 + 1) + 13 x2 ,
g3 = (1, 2), g4 = (1, 0)

1
3 x1

4
f (x) X
gi (x)
+
ui
xj
xj
i=1

i = 1, 2, 3, 4

i = 1, 2, 3, 4

+ 6(x2 2)), g1 = ( 29 x1 , 2x2 ), g2 = (5, 1),

(iii) (2p) x = (3, 1)T is a global minimum.

= 0, j = 1, 2

7.

(i) (2p) The dual problem is


max
u0

h(u)

where
h(u) =
s.t.

min 5x1 + 6x2 + 8x3 + 3x4 + u1 (5 2x1 3x2 x3 3x4 ) + u2 (8 4x1 x2 5x3 x4 )
xj {0, 1} j = 1, 2, 3, 4.(the Lagrangian subproblem)

(ii) (2p) (u1 , u2 ) = (1, 2)

x = (1, 0, 1, 1)

h(u) = 11

LBD = 11

(iii) (2p) Since x = (1, 0, 1, 1) is feasible to the constraints (1) and (2)
U BD 11 z 16.

z = 16 =

You might also like