Chair for Operations Management

Prof. Dr. Rainer Kolisch

Arcisstraße 21, 80333 München

Examination Management Science

Wednesday, August 3, 2016

Name: First Name: Matriculation Number:

Program: Semester: Signature:

Please stick to the following rules:

1. Fill in all required information in the box above.

2. Check the completeness of the exam: Front page (1 page), Part 1 – 4 (10 pages).

3. Use permanent ink. Do not use red ink or pencil.

4. Put you ID card and your study card in front of you so that we can check your identity.

5. Solve each part on separate pages. Write your name, your matriculation number, part and
number of the problem on each page, including additional pages. Number the pages with
the solution of each part consecutively.

6. Solve the problems in a readable and reproducible manner. In case of non-integer numbers,
perform calculations with two decimal places.

7. Do not remove the staple.

8. Non-programmable calculators are allowed only.

9. You have 60 minutes to write the exam.

Valuation (will be done by the examiner)

P
Part 1 2 3 4
Maximum Points 23 14 10 13 60
Obtained Points

218
Management Science Name:............................ Student number: ............................

Part 1 – Linear Programming (23 Points)

1.1 (8P) A chemical company wants to maximize the profit of its supply chain. In the supply
chain the two glues L1 and L2 are produced. For the production, the two pre-products
ammonia (A) and methanol (M) are required. Ammonia and methanol are produced on
the same resource which has a total production capacity of 500.000 kg. To produce 1 kg
of L1 0, 5 kg of ammonia and 0, 9 kg of methanol are needed. To produce 1 kg of L2 0,7
kg of ammonia and 0,4 kg of methanol are required. Ammonia and methanol which is not
used for the production of L1 and L2 can be sold on the market. Due to supply chain
contracts there are minimum quantities of L1 and L2 which have to be produced. Sales
prices, production costs and minimum quantities are given in Table 1.

Products A M L1 L2
Sales price (in monetary units) 10 20 30 35
Production costs (in monetary units) 5 8 14 16
Minimum quantities (in 1.000 kg) 0 0 5 7

Tabelle 1: Sales prices, production costs and minimum quantities

How many units shall the company produce in order to maximize profit? Formulate a
linear program and explain the variables, the objective function value and the constraints.

219
Management Science Name:............................ Student number: ............................

1.1 (12P) Consider the following maximization problem:

Max 1 · x1 + 1 · x2 = z

subject to the constraints

1 · x1 + 0 · x2 ≤ 4
0 · x1 + 1 · x2 ≤ 6
1
· x1 − 1 · x2 ≤ 0
4
1
· x1 + 1 · x2 ≤ 7
2

x1 , x2 ≥ 0

The start tableau is given by

BV Wert x1 x2 x3 x4 x5 x6

x3 4 1 0 1 0 0 0

x4 6 0 1 0 1 0 0

1
x5 0 4 −1 0 0 1 0

1
x6 7 2 1 0 0 0 1

−z 0 1 1 0 0 0 0

a) Which special case is given in the start tableau? What could be the effect of the
special case when applying the Simplex algorithm? (2P)

b) Starting with the tableu given above perform the next iteration of the primal or dual
Simplex in order to improve the objective function. Mark the pivot element in the
start tableau. Write the resulting tableau in the table below.
(6P)

220
Management Science Name:............................ Student number: ............................

BV Wert x1 x2 x3 x4 x5 x6

c) Find the optimal solution and the optimal objective function value graphically. Draw
the solution space in the coordinate system below. Number all corners of the solution
space which are visited by the Simplex algorithm. Start with the corner you obtained
in part a). Draw the linear slope of the objective function and all optimal points
derived. (4P)

x2

x1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

221
Management Science Name:............................ Student number: ............................

1.3 (3P) Consider the following linear program (LP)

Min 3 · x1 + 2 · x3 = z

subject to the constraints

2 · x1 + 4 · x2 ≥8
3 · x1 + 3 · x3 ≥ 2
2 · x1 + 5 · x2 + 1 · x3 = 10
x1 ≤0
x2 ≥0
x3 ∈ R.

Transform the linear program into its dual. (3P)

222
Management Science Name:............................ Student number: ............................

Part 2 – Integer Programming (14 Points)

Consider the following integer program:

Max z = 6 · x1 + 5 · x2

subject to

4 · x1 + 5 · x2 ≤ 25
6 · x1 + 4 · x2 ≤ 30
x1 ≥ 0, x2 ≥ 0 and x1 , x2 ∈ Z.

The LP relaxations of the subproblems are given in the following table:

Subproblem Constraint added to P0 Solution of relaxation

x1 x2 z
Pa 3, 75 2,14 32,14
Pb x1 ≤ 3 3 2,6 31
Pc x1 ≤ 3 x2 ≥ 3 2,5 3 30
Pd x1 ≤ 3 x2 ≤ 2 3 2 28
Pe x1 ≤ 2 x2 ≤ 3 2 3 27
Pf x1 ≥ 4 4 1,5 31,5
Pg x1 ≥ 4 x2 ≥ 2 infeasible
Ph x1 ≥ 4 x2 ≤ 1 4,33 1 31
Pi x1 ≤ 4 x2 ≤ 1 4 1 29
Pj x1 ≥ 5 x2 ≤ 1 5 0 30

223
Management Science Name:............................ Student number: ............................

a) Determine the optimal solution with branch–and–bound. Generate the branch–and–bound

tree according to LIFO with the larger character in the alphabet as tie breaker (i.e.
when choosing between Pa and Pz choose Pz first). Draw the branch–and–bound tree and
indicate in the tree which rule is used in order to fathom a subproblem. Indicate the order
in which subproblems are considered. (12P)

224
Management Science Name:............................ Student number: ............................

b) At what stage would the branch–and–bound procedure have been terminated if the ob-
jective function value could have been not more than 5% below the optimal objective
function value of the linear relaxation of the original problem? (1P)

c) Would the optimal solution have been found in less steps by applying the MUB–rule?
Provide an argument for your answer. (1P)

225
Management Science Name:............................ Student number: ............................

Part 3 – Graph Theory (10 Points)

Consider the following directed graph with minimum flows λi,j , maximum flows κi,j and a start
solution with flows xi,j .

[0, 4]
1 2
[0, 5] [0, 3]
[0, 3] 2 [1, 3]
2
q 4 s
[0, 3] [0, 2]
2 [0, 2]
2
[0, 2] [λi,j , κi,j ]
3 5 i j
xi,j

a) Determine the maximum flow by using the algorithm of Ford und Fulkerson in the table
below (9P).
Iterationen
Knoten Ausgangsfluss
q (+, ∞)

2 (4+ , 2)

3 (q + , 2)

4 (3+ , 2)

s (2+ , 2)
∆ 2
P
2

b) Determine the sets Vq und Vs , which will lead to a minimum cut. Draw the minimum cut
in the graph given above. (1P)

226
Management Science Name:............................ Student number: ............................

Part 4 – Decision Analysis (13 Points)

4.1 (4P) Consider the following decision matrix for a minimization problem:

s1 s2 s3
a1 x 1 x
a2 2 1 2
a3 5 −1 −2

a) Determine the efficient alternatives depending on x with x in the interval

[−10, −9, . . . , 9, 10]. Provide an argument for your answer. (2P)

b) Determine the maximum value for x for which a decision maker applying the Laplace
rule would choose alternative a1 . Provide a rationale for your answer. (2P)

4.2 (9P) A chemical company wants to introduce a new product to the market. In case of a
success (80% probability) a profit of 4 will be made. Otherwise, there will be a loss of 1.
Before introducing the product to the market, a market survey which costs x > 0 can
be undertaken. In case the market survey is positive (80% probability) the chance of a
market success is 98, 75%. If the market survey yields a negative result, the chance of a
market success is 5%.
The decision tree provides the three alternatives “Market introduction”, “Market survey”
and “No introduction to market” as well as the payoffs.

227
Management Science Name:............................ Student number: ............................

Market success
Market 4
0, 8
introduction

No market success
-1
0, 2
Market success

No market success
positiv

Market survey
Market success

negativ No market success

No market introduction
0

a) Complete the decision tree. (3P)

b) Use the roll back procedure in order to calculate the expected payoff of the three
alternatives depending on x. Indicate the optimal strategy in the decision tree for
x ≥ 1. (5P)
c) Using the decision tree, calculate the value of the market survey. (1P)

228