[Mat10] Globális Optimalizálás kurzussorozat

[Csilla Majoros]

*Meghívó*

Szeretettel várunk minden kedves érdeklődőt

*Julius Zilinskas *
(Vilnius University)

professzor úr *Globális Optimalizálás* mini kurzussorozatán.


*1. előadás*
2015. december 15-én kedden 10:00 - 11:00, BME H. épület IV. em, H45/a terem
Az előadás címe: *Global Optimization*

*Abstract:* Global optimization aims at minimization of a nonlinear
objective function when no assumptions on unimodality  are included into
formulation of the problem - many local minima may exist. We will start
this mini-course from introduction to global optimization and discussion of
deterministic covering methods partitioning the feasible region.

*2. előadás*
2015. december 15-én kedden 14:00 - 15:00, BME H. épület IV. em. H45/a terem
Az előadás címe:

*Simplicial Global Optimization **Abstract:* Although rectangular
partitions are used most often in global optimization, in this lecture we
will particularly discuss simplicial covering which has several advantages.
Applications benefiting from simplicial partitioning are examined in
detail: nonlinear least squares regression and pile placement in
grillage-type foundations.

*3. előadás*
2015. december 16-án szerdán 14:00 - 16:00, BME H. épület III. em. H306
terem
Az előadás címe: *Optimization Based Visualization of Multidimensional Data*

*Abstract:* The next lecture will focus on optimization based visualization
of multidimensional data. Multidimensional data visualization enables
exploratory data analysis involving heuristic abilities of human experts
when a set of objects is represented in a low-dimensional space. In this
lecture we consider one of the most popular approaches known as
multidimensional scaling, whose essential part is optimization of a
function possessing many optimization adverse properties. Application areas
of multidimensional scaling vary from psychometrics to pharmacology.

2015. december 17-én csütörtökön 14:00 - 16:00, BME H. épület III. em. H306
terem
Az előadás címe: *Non-convex Multi-objective Optimization*

*4. előadás*
*Abstract:* Real-world optimization problems usually involve more than one
criteria, what leads to solution of multi-objective optimization problems.
Multi-objective optimization discussed in this lecture is a research area
rich with various approaches. Many methods convert the multi-objective
optimization problem into a set of single-objective problems. Apart from
general disadvantages of such approaches, in non-convex multi-objective
optimization even the scalarized single-objective optimization problem is
not easily solved - global
optimization must be used. As we saw in previous lectures, branch and bound
method can be used to solve global optimization problems, and moreover it
is possible to develop multi-objective branch and bound algorithms which do
not require scalarization of the problems. In non-convex continuous
multi-objective optimization such a multi-objective branch and bound
algorithm can find close approximation of the Pareto front with predefined
accuracy and in discrete multi-objective optimization the exact
Pareto-optimal set can be found.

Literature:
R. Paulavičius, J. Žilinskas (2014) Simplicial Global Optimization.
Springer, ISBN 978-1-4614-9092-0. doi:10.1007/978-1-4614-9093-7.
[Abstracted/Indexed in ISI Web of Science (Book Citation Index),
SpringerLink, MathSciNet. Preview at Amazon.com, Google Books].
G. Dzemyda, O. Kurasova, J. Žilinskas (2013) Multidimensional Data
Visualization: Methods and Applications. Springer, ISBN 978-1-4419-0235-1.
doi:10.1007/978-1-4419-0236-8. [Abstracted/Indexed in SpringerLink,
MathSciNet. Preview at Amazon.com, Google Books].


Üdvözlettel,

Majoros Csilla
--------- következő rész ---------
Egy csatolt HTML állomány át lett konvertálva...
URL: <http://lists.math.bme.hu/pipermail/mat10/attachments/20151207/03679735/attachment.html>