350 rub
Journal Neurocomputers №5 for 2014 г.
Article in number:
Mathematical models of systems with interval representation of parameters on the basis of heterogeneous neural networks. Porous catalyst
Keywords: boundary value problem (BVP) interval parameter modeling artificial neural network training error functional global optimization
Authors:
A. N. Vasilyev - Dr.Sc. (Eng.), Professor, Department Higher Mathematics, Applied Mathematics and Mechanics Institute, St. Petersburg State Polytechnical University. E-mail: a.n.vasilyev@gmail.com
D. A. Tarkhov - Dr.Sc. (Eng.), Professor, Department Higher Mathematics, Applied Mathematics and Mechanics Institute, St. Petersburg State Polytechnical University. E-mail: dtarkhov@gmail.com
Abstract:
Neural networking technique with models based on differential equations is applied to known incorrect problems which solution by routine approaches has difficulties. An approximate solution to the problem is found as output of neural network with some prescribed architecture. Network weights are determined in the process of stepwise network training grounded on some error functional minimization in general case. We consider in the three-part composition the case when system parameters are given in some variation intervals. This paper - the first part of the composition - deals with BVP for ordinary differential equation with interval parameters. Construction of robust neural network model of processes in porous catalyst is cited as an example. The similar problem in the case of interval temperature conduction factor for both classical and nonclassical statements was solved via neurocomputing on growing neural networks in other two parts of the composition. Results of neurocomputing and some corresponding figures are given. Advantages of neural network approach and some possible generalizations are mentioned.
Pages: 3-7
References

  1. Vasil'ev A.N., Tarkhov D.A. Neyrosetevoe modelirovanie. Printsipy. Algoritmy. Prilozheniya. SPb.: Izd-vo SPbGPU. 2009. 528 s.
  2. Vasil'ev A.N. Neyrosetevoe modelirovanie v matematicheskoy fizike // Neyrokomp'yutery: razrabotka, primenenie. 2009. № 5. S. 25-38.
  3. Vasil'ev A.N., Tarkhov D.A. Neyrosetevaya metodologiya postroeniya priblizhennykh matematicheskikh modeley raspredelennykh sistem // Trudy nauchno-metod. seminara kafedry vyssh. matematiki. Vyp. 1. SPb.: Izd-vo Politekhn. un-ta. 2008. S. 115-170.
  4. Vasil'ev A.N. Postroenie priblizhennykh matematicheskikh modeley raspredelennykh sistem na osnove neyrosetevoy metodologii // Neyrokomp'yutery: razrabotka, primenenie. 2007. № 9. S. 103-116.
  5. Vasil'ev A.N., Tarkhov D.A., Osipov V.P. Unifitsirovannyy protsess modelirovaniya fiziko-tekhnicheskikh ob''ektov s raspredelennymi parametrami // Nauchno-tekhn. vedomosti SPbGPU. Fiz.-mat. nauki. 2010. № 3(104). S. 39-52.
  6. Tarkhov D.A. Neyronnye seti: modeli i algoritmy. Kn. 18. M.: Radiotekhnika. 2005. 256 s.
  7. Vasil'ev A.N. Tarkhov D.A. Neyrosetevoe reshenie zadachi o poristom katalizatore // Nauchno-tekhn. vedomosti SPbGPU. Fiz.-mat. nauki. 2008. № 6 (67). S. 110-113.
  8. Hlavacek V., Marek M., Kubicek M. Modelling of chemical reactors. Part X // Chem. Eng. Sci. 1968. V. 23. P. 1083-1097.
  9. Kubicek M., Hlavacek V. Solution of nonlinear boundary value problems. Part VIII // Chem. Eng. Sci. 1974. V. 29. P. 1695-1699.
  10. Dmitriev S.S., Kuznetsov Ye.B. Perenos tepla i massy v poristom katalizatore // Materialy VI Mezhdunar. konf. po neravnovesnym protsessam v soplakh i struyakh. NPNJ-2006. SPb. - M.: Vuzovskaya kniga. 2006. S. 159-160.
  11. Lahae M.E. Solution of systems of transcendental equations // Acad. R. Belg. Bull. Cl. Sci. 1948. V. 5. P. 805-822.
  12. Kuznetsov Ye.B. Nailuchshaya parametrizatsiya pri postroenii krivoy iteratsionnym metodom // Dokl. RAN. 2004. T. 396. № 6. S. 746-748.
  13. Na Ts. Vychislitel'nye metody resheniya prikladnykh granichnykh zadach. M.: Mir. 1982. 296 s.
  14. Vasil'ev A.N., Tarkhov D.A. Neyrosetevye podkhody k resheniyu kraevykh zadach v mnogomernykh sostavnykh oblastyakh // Izv. TRTU. 2004. № 9. S. 80-89.
  15. Vasil'ev A.N., Tarkhov D.A. Primenenie iskusstvennykh neyronnykh setey k modelirovaniyu mnogokomponentnykh sistem so svobodnoy granitsey // Izv. TRTU. 2004. № 9. S. 89-100.
  16. Vasil'ev A.N., Tarkhov D.A. Raschet teploobmena v sisteme «sosudy-tkani» na osnove neyronnykh setey // Neyro­komp'yutery: razrabotka, primenenie. 2006. № 7. S. 48-53.
  17. Vasil'ev A.N. Sravnitel'nyy analiz traditsionnogo i neyrosetevogo podkhodov k postroeniyu priblizhennoy modeli kalibratora peremennogo davleniya // Neyrokomp'yutery: razrabotka, primenenie. 2007. № 9. S. 14-23.
  18. Vasil'ev A.N., Tarkhov D.A. Evolyutsionnye algoritmy resheniya kraevykh zadach v oblastyakh, dopuskayushchikh dekompozitsiyu (NPNJ-2006) // Matematicheskoe modelirovanie. 2007. T. 19. № 12. S. 52-62.
  19. Vasil'ev A.N., Tarkhov D.A. Neyrosetevye podkhody k regulyarizatsii resheniya zadachi prodolzheniya temperaturnykh poley po dannym tochechnykh izmereniy // Neyrokomp'yutery: razrabotka, primenenie. 2010. № 7. S. 13-19.
  20. Vasil'ev A.N., Porubaev F.V., Tarkhov D.A. Neyrosetevoy podkhod k resheniyu nekorrektnykh zadach teploperenosa // Nauchno-tekhn. vedomosti SPbGPU. Informatika. Telekommunikatsii. Upravlenie. 2011. № 1(115). S. 133-142.
  21. Vasil'ev A.N., Tarkhov D.A. Postroenie priblizhennykh neyrosetevykh modeley po raznorodnym dannym // Matematicheskoe modelirovanie. 2007. T. 19. № 12. S. 43-51.
  22. Samarskiy A.A., Vabishchevich P.N. Chislennye metody resheniya obratnykh zadach matematicheskoy fiziki. M.: Yedito­rial URSS. 2004. 480 s.
  23. Vasil'ev A.N., Tarkhov D.A. Matematicheskie modeli sistem s interval'no zadannymi parametrami na osnove geterogennykh neyronnykh setey. Prodolzhenie temperaturnogo polya ? klassicheskaya postanovka zadachi // Neyrokomp'yutery: razrabotka, primenenie. 2012. № 11. S. 56-59.
  24. Vasil'ev A.N., Tarkhov D.A. Matematicheskie modeli sistem s interval'no zadannymi parametrami na osnove geterogennykh neyronnykh setey. Prodolzhenie temperaturnogo polya ? neklassicheskaya postanovka zadachi // Neyrokomp'yutery: razrabotka, primenenie. 2012. № 6.