IATOM - ENAC – Combinatorial Optimization

Exercises IV

Classical combinatorial optimization problems

Exercise 13: Shortest path problem

In graph theory, The shortest path problem is the problem of finding a path between
two vertices (or nodes) in a graph such that the sum of the weights of its constituent edges is
minimized.
1) Model the shortest path problem from node A to node F in the graph below, as an
integer linear program.
2) Apply an appropriate algorithm to solve this problem.

B
8
1
5 E
1
4 1
A C
3
3 2
F
5
D

Exercise 14: An Air Transportation problem

A freight airline has to carry some product from several stock centers to several
distribution centers, linked together by the air network represented on the following graph.
In this graph, the distances are represented on arcs and we assume that the cost
transportation from one node to another one is linearly dependant on the distance of the
used path. Stock centers are located at nodes 1 and 2 and distribution centers are located
at nodes 3, 4 and 5.
Which paths can use the freight carrier so as to minimise the cost transportation?

2 4
1 6 10
1
3 3 2 3 3
3
2 7 4 11 1 9
4 2
1 3 6 3 2
2
8 3 5 4
2 2 2

1) Which kind of optimization problem is that?

2) Use an appropriate algorithm to solve the problem. Detail each step of the algorithm.
Exercise 15: Knapsack problem
Let us consider a knapsack with a maximum load of 70 kg.
Let us consider a set of 4 objects with the following features:

object number Value (v) weight (w)

1 140 20
2 120 30
3 100 20
4 60 10

The knapsack problem consists in selecting a sub-set of these objects (each object cannot be
selected more than once) to be packed in the knapsack and that maximize the total value of
the knapsack.
1) Model this problem
2) Propose a method to solve this problem. Is it an exact method?
3) Solve it with a branch-and-bound. Which method is used to solve each continuous
relaxation? Is it an exact method?

Exercise 16: Ground handling activities scheduling problem

Ground activities around an aircraft (for passenger traffic) during a stopover can be
summarized to the following : aircraft parking (A), passenger deboarding (B), baggage
unloading (C), sewage draining and water supply (D), passenger cabin cleaning (E),
refueling (F), catering (G), baggage loading (H), passenger boarding (I), door closing and
push-back (J). Let us suppose that the expected duration of these activities and their
sequencing relationships are given in the following table.

Activity A B C D E F G H I J
Duration 5 15 15 10 15 20 15 15 30 5
Previous - A A B B C,D E F G,H I,H
activities

We want to know what is the minimum duration of such a stopover.

1) Model this problem with a linear program.

4) Give a graph representation of this problem.
5) Using this graph, compute the minimum duration of the stopover and identify which
activities cannot be delayed if one wants to respect this minimum duration (critical
activities).
6) Compute the earliest start time, latest start time of each task, in order to respect the
minimum duration.
7) Give a Gantt diagram of the schedule.
Exercise 17: Maximum flow problem
The maximum flow problem consists in finding a feasible flow through a single-
source, single-sink flow network that is maximum, while satisfying the capacity of each arc of
the network.

1) Model the maximum flow problem from source s to sink t in the network below, where the
arcs are valuated with their capacity, as an integer linear program.
2) How can we solve this problem?

15
A D
10
20 25

10 5
s B E t
20
10
35 10 20
5
C F
15

Exercise 18:
Let us consider the following graph where each arc is associated with its length.

1
2 6
4
3
1 2
3
1 1 4
6
8 1
2
1
5
3
4
7

We want to find paths from node 1 to node 4 with no common arc (no-adjacent paths).
3) Show that this problem comes down to a maximum flow problem in a particular
network.
4) Solve the problem.

Exercise 19: Air traffic capacity

 Let us consider the following air traffic network where each arc is associated with its
capacity (number of aircraft per time unit).

2 2 3
2 4 7 11
5
3 2 3 2 5
4
1 4 5 8 12

3 3 3 3
5 6
3 6 9 10
2 4 2

8) Determine the maximum number of aircraft that can go across this network per time
unit between node 1 and node 12.
9) Answer the same question when node 5 is forbidden.
10) Answer the same question when arc (5,8) is forbidden.

Exercice 20:
Formulate (do not solve!) the following problem as a Mixed Integer Linear
Programming problem.
We produce two fresh juices: grape juice and apple juice. For producing them we can
use two possible industrial processes: process 1 and process 2. the cost of each process
depends upon the quantity of juice produced as follows:
– it costs G euros per liter of grape juice produced with process 1.
– it costs A euros per liter of apple juice produced with process 1.
– process 2 involves a fixed cost of K euros dedicated to buy a machine required to
process any amount of juice (the same machine can be used for both juices).
– Once the machine is bought, the remaning cost of production using process 2 is C
euros per liter of juice produced.
We can produce up to 3,000 liters of grape juice and we cannot produce more than a total of
3,500 liters of juice. We can sell our production to a supermarket that wants to buy 1,000 liters
of each juice at the cost of 2 euros per liter. The supermarket accepts to buy more than 1,000
liters of grape juice but at the reduced price of 1 euro per extra liter. However, it will not buy
any extra liter of apple juice.

Consider that the constants G, A, K and C are given data, (all positive). The problem is to
decide how many liters of each juice we should produce with each process, in order to
minimize the overall costs (overall costs = production costs-revenues).