350 rub
Journal Neurocomputers №2 for 2014 г.
Article in number:
Neuronetwork algorithms of mathematical modeling processes and the phenomena in difficult systems on computer systems of the ultrahigh productivities
Keywords: High-Performance Computing (HPC) modeling artificial neural network training error functional global optimization evolutional algorithm
Authors:
A.N. Vasiliev - Dr.Sc. (Eng.), Professor, St.-Petersburg State Polytechnical University. E-mail: a.n.vasilyev@gmail.com
D.A. Tarkhov - Dr.Sc. (Eng.), St.-Petersburg State Polytechnical University. E-mail: dtarkhov@gmail.com
Abstract:
The article is devoted to the authors developed neural network approaches to solving the fundamental problems that arise in the solution of problems of complex system mathematical modeling on supercomputers and ultra-high-performance computer systems. We discuss algorithms to implement these approaches. These algorithms are resistant to errors in input data and allow of some efficient parallel and distributed realization, versions of which are discussed in the paper. They can be adapted to newly arriving data on the simulated system, and they can rebuild model type and structure during model construction and during operation of the modeled object. The resulting models are reusable in similar tasks, and (or) in the subsequent stages of research, and they can be distributed as separate products.
Pages: 7-15
References

  1. Il'in V.P. E'kzaflopsy' protiv matematicheskogo modelirovaniya //«Otkry'ty'e sistemy'» 2013. № 5.
  2. Gergel' V.P., Linev A.V. Problemy' i perspektivy' dostizheniya e'kzaflopnogo urovnya proizvoditel'nosti superkomp'yuterny'x sistem //Informaczionny'e texnologii. Vestnik Nizhegorodskogo universiteta im. N.I. Lobachevskogo. 2012. № 3(1). S.189-198.
  3. Gorbachenko V.I. Nejrokomp'yutery' v peshenii kraevy'x zadach teorii polya. Kn.10. M.: Radiotexnika. 2003. 336 s.
  4. Vasil'ev A.N., Tarxov D.A. Nejrosetevoe modelirovanie. Princzipy'. Algoritmy'. Prilozheniya. SPb.: Izd-vo SPbGPU. 2009. 528 s.
  5. Tarxov D.A. Nejronny'e seti: modelirovanie i algoritmy'. Kn. 18. M.: Radiotexnika. 2005. 256 s.
  6. Gorbachenko V.I., Artyuxina E.V. Bessetochny'e metody' i ix realizacziya na radial'no-bazisny'x nejronny'x setyax// Nejrokomp'yutery': razrabotka, primenenie. 2010.  № 11. S.4-10.
  7. Vasil'ev A.N., Osipov V.P., Tarxov D.A. Unificzirovanny'j proczess postroeniya ierarxii nejrosetevy'x modelej i problema adekvatnosti v matematicheskom modelirovanii // Nejrokomp'yutery': razrabotka, primenenie. 2012. № 7. S.20-28.
  8. Tarxov D.A. Nejronny'e seti kak sredstvo matematicheskogo modelirovaniya // Nejrokomp'yutery': razrabotka, primenenie. 2006. № 2. S.3-48.
  9. Manoj K., Neha Y. Multilayer perceptrons and radial basis function neural network methods for the solution of differential equations: A survey// Computers and Mathematics with Applications. 2011. V. 62. Issue 10. P.3796-3811.
  10. Vasil'ev A.N., Tarxov D.A. Parametricheskie nejrosetevy'e modeli klassicheskix i neklassicheskix zadach dlya uravneniya teploprovodnosti Nauchno-texnicheskie vedomosti SPbGPU. Fiz.-mat.nauki. SPb.: Izd-vo SPbGPU. 2012. S.136-145.
  11. Vasil'ev A.N., Tarxov D.A. Matematicheskie modeli sistem s interval'no zadanny'mi parametrami na osnove geterogenny'x nejronny'x setej. Prodolzhenie temperaturnogo polya ? klassicheskaya postanovka zadachi // Nejrokomp'yutery': razrabotka, primenenie. 2012. № 11. S.56-59.
  12. Vasil'ev A.N., Porubaev F.V., Tarxov D.A. Nejrosetevoe reshenie dvumernoj obratnoj zadachi teploperenosa s tochechny'mi danny'mi izmerenij // Nejrokomp'yutery': razrabotka, primenenie. 2011. № 6. S.38-44.
  13. Tarxov D.A. Netradiczionny'e geneticheskie algoritmy' dekompoziczii i raspredeleniya pri reshenii zadach matematicheskoj fiziki s pomoshh'yu nejronny'x setej // Nejrokomp'yutery': razrabotka, primenenie. 2006. № 7. S.40-47.
  14. Tarxov D.A. Nekotory'e podxody' k postroeniyu raspredelenny'x nejrosetevy'x modelej // Trudy' 5-j mezhdunar. nauchno-texn. konf. «Komp'yuternoe modelirovanie 2004». SPb. 2004. S.170-171. 
  15. Tarxov D.A.  Raspredelenny'e vy'chisleniya i nejronny'e seti // Sb. dokl. mezhdunar. konf. Po myagkim vy'chisleniyam i izmereniyam SCM-2004. SPb. 2004. T. 1. S.295-298.
  16. Intellektualizacziya slozhny'x sistem // Informaczionno-izmeritel'ny'e i upravlyayushhie sistemy'. 2009. T. 7. № 3. S. 1-92.
  17. Rozhnov A.V. Nekotory'e problemny'e voprosy' sistemnoj integraczii napravlenij nauchnoj deyatel'nosti v zadachax nejrokomp'yutinga // Nejrokomp'yutery': razrabotka, primenenie. 2014. № 1. S. 3-9.