第一个网页(共2个)
点击打开链接 http://www1.osu.cz/~tvrdik/?page_id=18
Algorithms of global optimization and their applications (English)
Global Optimization, Evolutionary Algorithms and their Application to Computational Statistics
Stochastic algorithms of global optimization
In 1993 we started to deal with stochastic algorithms of global optimization. We modified the Controlled Random Search algorithm (Price 1976) by randomizing the reflection in simplex. The modified CRS algorithm was applied to estimating parameters of non-linear regression models and also in the shape optimization problem. For the testing of the stochastic algorithms a set of fourteen difficult non-linear regression tasks was collected. This collection can be downloaded here, the description of the tasks and references are in the file nlmod14.pdf (zip), data in text format are in the file models.zip.
Differential Evolution:
Competitive Diferential Evolution for Constrained Problems – CEC 2010
Source code in Matlab download readme.txt (source code zip)
Adaptive Diferential Evolution Algorithms, Handbook of Optimization, 2012
Source code in Matlab or C download readmeHO.txt (source code zip)
Hybrid DE for Optimal Clustering 2015
Manuscript of paper download asoc2014sub_tvkr
Matlab source code readmeClust download clustan_b6e6rl_source
Matlab Program Library for Box-constrained Continuous Problems (including a new adaptive stochastic algorithm for the estimation of parameters
in nonlinear regression models) here
Selected papers:
Tvrdík, J., Estimation of Regression Parameters by Controlled Random Search Algorithm (in Czech), In: Antoch, J. and Dohnal, G. (eds), Proceedings of ROBUST 94, 153-159, JČMF Praha, 1994, ISBN 80-7015-492-6
Křivý, I., Tvrdík, J., The Controlled Random Search Algorithm in Optimizing of Regression Models, Computational Statistics and Data Analysis, 20, 229-234, 1995
Haslinger J., Jedelský D., Kozubek T., Tvrdík J., Genetic and Random Search Methods in Optimal Shape Design Problems, Journal of Global Optimizations 16, 109-131, 2000
Křivý I., Tvrdík J., Stochastic Algorithms in Estimating Regression Models, COMPSTAT 1996, Proceedings in Computational Statistics (ed. A. Prat), 325-330, Physica Verlag, 1996
Tvrdík, J., Křivý, I., Simple Evolutionary Heuristics for Global Optimization, Computational Statistics and Data Analysis, 30, 345-352, 1999
download allmut99.pdf (zip)
Křivý I., Tvrdík J., Krpec R.: Stochastic Algorithms in Nonlinear Regression, Comput. Statist, Data Anal. 33, 278-290, 2000
Mišík, L., Tvrdík, J., Křivý, I. : On Convergence of a Class of Stochastic Algorithms, In: Proceedings of ROBUST 2000 (J.Antoch a G. Dohnal eds), 198-209, JČMF Praha, 2001, ISBN 80-7015-792-5
download rob00new.pdf (zip)
Tvrdík, J., Křivý, I., Mišík, L., Evolutionary Algorithm with Competing Heuristics, In: Ošmera, P. (ed.) Proceedings of MENDEL 2001, 7th International Conference on Soft Computing, 58-64, Technical University, Brno, 2001,
download mendel01.pdf (zip)
Tvrdík, J., Mišík, L., Křivý, I., Competing Heuristics in Evolutionary Algorithms, 2nd Euro-International Symposium on Computational Intelligence June 16 – 19, 2002, Košice, Slovakia, published In SINCAK et al. Intelligent Technologies-Theory and Applications. Amsterdam, The Nethelands: IOS Press, 2002. 159-165.
download kos02cln.pdf (zip)
Tvrdík, J., Křivý, I., Mišík, L., Evolutionary Algorithms with Competing Heuristics in Computational Statistics, COMPSTAT 2002, Proceedings in Computational Statistics (ed. Haerdle W., Roenz B.), 349-354, Physica Verlag, Heidelberg
download comp02.pdf (zip)
Tvrdík, J., Generalized controlled random search and competing heuristics. In MENDEL 2004, 10th Internetional Conference on Soft Computing (Matoušek R. and Ošmera P. eds). University of Technology, Brno, 2004. 228-233.
download mendel04.pdf
Tvrdík, J. Competition and Cooperation in Evolutionary Algorithms: A Comparative Study. In MENDEL 2005, 11th International Coference on Soft Computing. Brno : University of Technology, Brno, 2005. 108-113.
download mendel05.pdf extended CD version men05_CD.pdf
Tvrdík, J., Křivý, I., Mišík, L., Adaptive population-based search: application to estimation of nonlinear regression parameters. Computational Statistics and Data Analysis 52(2), 713-724 (2007)
download preprint CSDA-06SAS03e.pdf
Tvrdík, J. Competitive Differential Evolution. In MENDEL 2006, 12th International Coference on Soft Computing. Brno : University of Technology, Brno, 2006. 7-12.
download mendel06.pdf extended CD version mendel06_CD.pdf
Tvrdík, J. Differential Evolution with Competitive Setting of its Control Parameters. TASK Quarterly 11, 169-179 (2007)
download reprint task07.pdf
Tvrdík, J. Adaptation in Differential Evolution: A Numerical Comparison. APPL SOFT COMPUT. 2009, 9, 1149-1155. ASOC2009_reprint.pdf
Tvrdík, J. Self-adaptive Variants of Differential Evolution with Exponential Crossover. Analele Universitatii de Vest, Timisoara.Seria Matematica-Informatica. 2009, 47, 151-168.
download reprint analele09.pdf
TVRDÍK, J., POLÁKOVÁ, R. Competitive Differential Evolution for Constrained Problems. 2010 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC). IEEE, 2010. s. 1632-1639.
POLÁKOVÁ, R., TVRDÍK, J. Various Mutation Strategies in Enhanced Competitive Differential Evolution for Constrained Optimization. 2011 IEEE Symposium on Differential Evolution. IEEE, 2011. s. 17-24. [2011-04-11]. ISBN 978-1-61284-070-3
TVRDÍK, J., KŘIVÝ, I. Hybrid Adaptive Differential Evolution in Partitional Clustering. MENDEL 2011 17th International Conference on Soft Computing. Brno: University of Technology, 2011. s. 1-8. [2011-06-15]. ISBN 978-80-214-4302-0 download preprint mendel11tvkr.pdf
第二个网页(共2个)Self-adaptive Algorithms for Global Optimization – MATLAB and C++ Library Including Parallel Implementation for Many Processors
Introduction
The Matlab program library has been established in the frame of the project201/05/0284 of the Czech Grant Agency at University of Ostrava in 2006.The extension by C++ version comes from next period and the parallel versionof C++ library is the result of the project supported by the European RegionalDevelopment Fund in the IT4Innovations Centre of Excellence project(CZ.1.05/1.1.00/02.0070)
Project team:
- Josef Tvrdík, Department of Comp. Science, University of Ostrava,
- Viktor Pavliska, Institute for Research and Applications of Fuzzy Modeling,University of Ostrava,
- Radek Valášek, Institute for Research and Applications of Fuzzy Modeling,University of Ostrava,
- Radka Poláková, Institute for Research and Applications of Fuzzy Modeling,University of Ostrava,
- Hashim Habiballa, Department of Comp. Science, University of Ostrava,
The aim of this library is to make new self-adaptive stochastic algorithmsaccessible for potential users and facilitate the use of the self-adaptivealgorithms in practical problems. The self-adaptation of the algorithms meansthat the user need not set up their control parameters by preliminarytrial-and-error numerical experiment.
The library includes source codes of algorithms:
- adaptive controlled random search (CRS) for the estimation of non-linear regression parameters (Matlab, non-parallel)
- competitive-adaptive differential evolution (DE) for boundary-constrained global optimization (Matlab, non-parallel)
- competitive-adaptive differential evolution (DE) for boundary-constrained global optimization (C++, non-parallel and/or parallel)
These programs are free software; you can redistribute it and/or modify itunder the terms of the GNU General Public License as published by the FreeSoftware Foundation; either version 2 of the License, or (at your option) anylater version. The programs are distributed in the hope that they will beuseful, but WITHOUT ANY WARRANTY; without even the implied warranty ofMERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General PublicLicense for more details, http://www.fsf.org/.
If you use a routine of this library, please, cite in your paper or report thisreference:
Tvrdík J., Pavliska V., Valášek, R., Poláková, R., Habiballa, H.(2014) Self-adaptive Algorithms forGlobal Optimization – MATLAB and C++ Library Including Parallel Implementationfor Many Processors, irafm.osu.cz/sago
Contact address for Matlab routines: josef.tvrdik@osu.cz, radka.polakova@osu.cz
Contact address for C++ routines: viktor.pavliska@osu.cz, radek.valasek@osu.cz