Introduction to Operations Research · 2022 - 2023

Introduction to Operations Research

2022 - 2023

Assignment 1
(10% of course grade)

Due Date: 7 December 2022, 11:00

1. The assignment is to be completed by the assignment team you created in Brightspace. No changes
to the teams are allowed at this point.

2. Hand in your answers before the deadline via Brightspace (one submission per team). You must
turn in a PDF file; points will be deducted if you turn in any other file formats. You do not need
to turn in a paper copy.

3. Put your group number in the name of the file, so that the name of the file becomes “Assign-
ment1_Group[group number].pdf”. Please do not use any compression software (such as ZIP, RAR
or ARJ).

4. Put your names and student numbers and your group number at the top of the first page.

5. This assignment will be graded out of 100, and will constitute 10% of your overall course grade.

Question 1 (50 points)

(a) Recall the standard primal-dual pair as given below:

Primal Problem Dual Problem

max Z = c T x min W = bT y
s.t. Ax ≤ b s.t. AT y ≥ c
x≥0 y≥0

By only using the definitions, properties, and theorems covered in this course, prove the following
statement: If the primal has an unbounded objective value, then the dual is infeasible. Similarly, if
the dual has an unbounded objective value, then the primal is infeasible.

(b) Consider the same primal-dual pair given in part (a), where x ∈ Rn and y ∈ Rm . Let ai
and Aj denote the ith row and the j th column of the constraint matrix A ∈ Rm×n , respectively.
Complementary slackness theorem states that

(ATj y∗ − cj )xj∗ = 0 for all j = 1, . . . , n, and

y∗i (bi − ai x ∗ ) = 0 for all i = 1, . . . m

if and only if x ∗ is an optimal solution to the primal and y∗ is an optimal solution to the dual.
By only using the definitions, properties, and theorems covered in this course, prove the comple-
mentary slackness theorem.
Note: When proving the "only if" statement (i.e., if the complementary slackness conditions hold for
x ∗ and y∗ , then x ∗ is primal optimal and y∗ is dual optimal), you can assume that x ∗ is primal
feasible and y∗ is dual feasible.

1
Introduction to Operations Research · 2022 - 2023

(c) Consider the following problem that is observed in a manufacturing plant:

max Z = 2x1 + 4x2 − x3

s.t. 3x1 + 2x2 − x3 ≤ 20
x1 − 5x2 + 3x3 ≤ 10
7x1 + 10x2 − 4x3 ≤ 12
x1 , x2 , x3 ≥ 0

Let x4 , x5 , x6 denote the slack variables for the respective constraints. After applying the Simplex
Method, the final (optimal) Simplex Tableau is given below.

Coefficient of: Right-hand

Z
x1 x2 x3 x4 x5 x6 Side
1 3.5 0 0 0 0.6 0.7 14.4
0 2.5 0 0 1 0.2 -0.1 20.8
0 4.5 0 1 0 1 0.5 16
0 2.5 1 0 0 0.4 0.3 7.6

Plant management is investigating options for making changes to the physical and the financial
characteristics of the system. The envisioned changes will affect the objective coefficient of x1 (c1 ),
the objective coefficient of x3 (c3 ), the second constraint coefficient of x1 (a21 ), and the right-hand
side of the second constraint (b2 ).
The envisioned changes in the system are interdependent. If a21 is changed by ∆, c3 will also
change by ∆, c1 will change by 4.5∆, and b2 will change by −4∆.
For the optimal basis given in the final Simplex Tableau to remain optimal, how much can the four
parameters c1 , c3 , a21 , b2 change? In another words, for which values of ∆ does the current optimal
basis remain optimal?

Question 2 (30 points)

At a specific department in a factory, a set of n jobs need to be processed. The same equipment is
used for each of the jobs, therefore simultaneous processing of multiple jobs is not possible. When the
processing of a job starts it needs to be finished completely before a new job can be processed.

For each of the jobs i there is a release date ri , a processing time pi , and a due date di . The processing
of a job can only start at or after its release date. If the processing of a job is finished later than its due
date it is considered late. For each job i, delay is defined as how much later than the due date the job
is finished.

There are also some operational precedence constraints that need to be satisfied. Given that p, q, r, s, t, u ≤
n denote distinct jobs,
• Job p must precede job q.
• If job r precedes job s, then job t must precede job u.
• Either job q precedes job s or job q precedes job u (or both).
Formulate this problem as a(n) (integer) linear programming problem, with the objective of minimizing
the maximum delay. Clearly define your decision variables and parameters.

2
Introduction to Operations Research · 2022 - 2023

Question 3 (20 points)

Ilke has made an agreement with her friends to catch up on a few episodes of a TV show so that they can
have a chat about it on Sunday. She must watch 10 episodes of the show on Friday and Saturday. Due
to her planning requirements, she must watch at least 3 episodes on Friday and at least 5 episodes on
Saturday. Moreover, she can watch up to 3 additional episodes (on top of the 10 that she must watch).
For each additional episodes that she watches, she will gain a popularity boost in her friend group, which
in her opinion amounts to b euros.

Ilke is subscribed to two streaming services that stream the episodes of this TV show: RUGflix and
RUGtube. RUGflix allows watching at most 4 episodes per day, and RUGtube allows at most 5 episodes
per day. Ilke’s subscription to both streaming services requires that she watches at least one episode on
each service on Friday. There is no such requirement for Saturday.

Ilke’s internet connection is metered, so watching an episode of the TV show on different days using
different streaming services has different costs. Watching an episode on RUGflix costs c11 and c12 euros
on Friday and Saturday, respectively; and watching an episode on RUGtube costs c21 and c22 euros on
Friday and Saturday, respectively.

Show how to model this problem in the standard form of the transportation problem. Since the mathe-
matical formulation is just the standard format as described in the lecture, you need not write down the
objective function and constraints. Rather, answer this question by clearly specifying
(a) Sources and destinations,
(b) Supply of each source and demand of each destination, and
(c) Transport costs between each (source, destination) pair.
Present the transportation costs in a matrix. Make sure the indexing in this matrix matches the indexing
you use when specifying sources and destinations. Briefly explain your choices.