You are on page 1of 2

Process Optimization

Tutorial 0

Summer Semester 2012


Department of Chemical Engineering
Malte Behrens, Sergio Lucia
Chair of Process Dynamics and Operations,
Prof. Sebastian Engell

General introduction and classes of optimization problems


Solving optimization problems is an invertible procedure in nearly every field of applications.
Many of them, especially in the field of engineering, can be formulated as mathematical
problems and can thereby be solved numerically or analytically. The broad range of
applications in which optimization problems occur led to a broad range of numerical solvers
and solution techniques. A classification of the concrete problem, at least from the
mathematical point of view, is therefore a very important step for the users of numerical
algorithms, optimization software and tools. This tutorial gives you a short introduction.

1. Questions
a) List the three basic elements that define an optimization problem.
b) Assign these three elements to the corresponding aspects of the following example
problems:
1. The producers of a product (on the left side of the figure) want to decrease
their shipping costs for distributing their product to the customers (on the
right side). The producers have a specific amount of product and the
customers have a specific demand. The costs for shipping depend on the
transport way and linearly on the amount of product ( costs for one
specific way: factor for the transport way multiplied with the amount of
product).
Storage A
Customer a
Storage B

Customer b

2. Similar to problem one, but there are two more products which can only be
transported in packages of 5 kg.
3. A model which describes the influence of the temperature T on the
viscosity of a liquid was developed. It has two unknown parameters
and which should be experimentally determined. A fixed set of
measurements are available (at Ti the viscosity i was measured; i=1, 2,..n).
In order to determine the parameters, the difference between the model and

1/2

Process Optimization

Tutorial 0

the measured values should be minimized. It is known, that and are


also coupled by linear and nonlinear inequalities g.
c) Which graphical representation could qualitatively belong to which problem? Label
the figures.

z2

x2

x1

z1

x2

x1
d) Classify these figures in terms of convexity of the feasible region if the equality
constraints are neglected (what is convexity?).
e) Which definition belongs to which problem?

minn c T x
xR

s.t. Ax b = 0
Cx d 0

minn c T x + 0.5 x T Bx
xR

s.t. g ( x) 0

minn f ( x, z )
xR
zZ m

s.t. h( x, z ) = 0
g ( x, z ) 0
where f(x,z), h(x,z) and g(x,z) are
affine functions of x and z

2/2

You might also like