Vous êtes sur la page 1sur 2

FIRST NAME

LAST NAME
REGISTRATION NUMBER
Optimization methods and algorithms - 01OUV
13/02/2013
1. Write the Linear Programming formulation of the Min Cost Flow Problem of the
following network (use the given data)
1
2
3
4
5
6
3
6
2
2
2
3
10 4
6
Adding the constraint that only one between the arcs (2,5) and (2,4) can have
non-zero ow in an optimal solution.
Which is the structure of any basic solution of the Min Cost Flow Problem? Give
an example on the given network.
2. Consider the following maximization problem
max z =

iI
c
i
x
i

iI
a
1
i
x
i
b
1

iI
a
2
i
x
i
= b
2
x
i
Z
+
, i I.
We want to solve it by means of a Branch & Bound method with continuous
relaxation. Let us suppose that, at a generic step, we have the search tree in the
picture, where, for each leaf node, the values [LB, UB, V
c
=val ] represent lower
bound, upper bound, and value in the continuous relaxation of the non-integral
variable chosen for branching, respectively.
Which is the next node we explore if we use the Best First policy?
Which are the nodes that we can close according to the non-improving fatho-
ming criterion?
Which are the new nodes generated by a binary branching scheme and the
associated additional constraints (write the new search tree after the bran-
ching)?
Justify your answers.
0
1 2
3 4 5 6
[9, 11, x
3
= 1.2] [10, 14, x
1
= 2.5] [12, 13, x
2
= 1.9] [11, 15, x
3
= 2.1]
x
2
= 0 x
2
1
x
3
2 x
3
3 x
4
5 x
4
6
3. We are solving a 0/1 Knapsack problem by means of a Genetic Algorithm. We
have chosen the following elements in the population as parents for the crossover
operator:
p1 = (231|8467|59) p2 = (324|1875|96),
where the chromosomes are coded such that any gene is an integer and each value
represents the item to be loaded in our knapsack (the instance contains 9 items).
Write the osprings we obtain by applying the Order Crossover (cut points after
genes 3 and 7).
4. Give the denition of P, NP and NP Complete computational complexity
classes.

Vous aimerez peut-être aussi