Thesis topic proposal
Tibor Csendes
New algorithmic tools for reliable optimization


Institute: University of Szeged
computer sciences
PhD School in Computer Science

Thesis supervisor: Tibor Csendes
Location of studies (in Hungarian): SZTE
Abbreviation of location of studies: SZTE

Description of the research topic:

The aim of the planned research is to clear the applicability of the new algorithmic tools developed in the recent years for reliable optimization. These are the directed acyclic graphs (DAGs) applied for better inclusion functions, constraint propagation, automatic differentiation, slope enclosures, and the new "cloud" paradigm for the representation of uncertainty. It is to be cleared how to apply them jointly in an effective and efficient way.
The PhD student is asked to set up a proper data structure suitable to allow a good implementation (operator overloading, new data types etc.). The work to be done is based by the recent collaboration with the University of Vienna. The investigation involves computational testing, algorithm development, and theoretical research as well.

The literature is again mostly in English:

H. Schichl and A. Neumaier, Interval Analysis on Directed Acyclic Graphs for Global Optimization, J. Global Optimization 33 (2005), 541-562: {http://www.mat.univie.ac.at/~neum/ms/dag.pdf}

A. Neumaier, Clouds, fuzzy sets and probability intervals, Reliable Computing 10 (2004), 249-272: {http://www.mat.univie.ac.at/~neum/ms/cloud.pdf}

and the COCONUT project page: {http://www.mat.univie.ac.at/users/neum/public_html/glopt/coconut/}

Two further reading on general optimization basics:

Bazara, M.S., H.N. Sherali and C.M. Shetty: Nonlinear Programming, John Wiley and Sons, New York, 1993

Horst, R. and P.M. Pardalos (eds.): Handbook of Global Optimization. Kluwer, Dordrecht, 1995

Number of students who can be accepted: 2

Deadline for application: 2017-08-31

All rights reserved © 2007, Hungarian Doctoral Council. Doctoral Council registration number at commissioner for data protection: 02003/0001. Program version: 1.2357 ( 2017. V. 15. )