NEURAL NETWORK MODEL FOR IDENTIFICATION OF MATERIAL CREEP CURVES USING CUDA TECHNOLOGIES

2019;
: 11-16
https://doi.org/10.23939/ujit2019.01.011
Received: October 25, 2019
Accepted: November 20, 2019
1
Lviv Polytechnic National University, Lviv, Ukraine
2
Ukrainian National Forestry University
3
Ukrainian National Forestry University
4
Ukrainian National Forestry University

This pa­per addres­ses the prob­lem of iden­tif­ying rhe­olo­gi­cal pa­ra­me­ters of wo­od using ar­ti­fi­ci­al neu­ral net­works with pa­ral­lel le­ar­ning al­go­rithm using Python prog­ram­ming lan­gua­ge, Cha­iner fra­me­work and CU­DA techno­logy. An in­tel­li­gent system for iden­ti­fi­ca­ti­on of rhe­olo­gi­cal pa­ra­me­ters of wo­od has be­en de­ve­lo­ped. The system cre­ated con­ta­ins the most user-fri­endly in­ter­fa­ce, all the ne­ces­sary set of to­ols for au­to­ma­ti­on of the pro­cess of vis­ua­li­za­ti­on and analysis of da­ta. In the pro­cess of cre­ati­on of the in­tel­lec­tu­al system, the fol­lo­wing tasks we­re en­vi­sa­ged: to carry out the analysis of ar­ti­fi­ci­al in­tel­li­gen­ce systems and the analysis of tra­ining of ar­ti­fi­ci­al neu­ral net­works, in par­ti­cu­lar mul­ti­la­yer neu­ral net­works of di­rect pro­pa­ga­ti­on, re­cur­rent neu­ral net­works and the Ko­ho­nen neu­ral net­work; exa­mi­ne the struc­tu­re of the Cha­iner fra­me­work and its in­te­rac­ti­on with CU­DA; to con­duct exis­ting clo­ud techno­lo­gi­es to ac­complish the task; to con­duct the analysis of al­go­rithms of stu­di­es of ar­ti­fi­ci­al neu­ron net­works, the­ir mat­he­ma­ti­cal pro­vi­ding; to imple­ment pa­ral­le­li­za­ti­on of le­ar­ning al­go­rithms and to de­ve­lop the ne­ces­sary softwa­re. Using Cha­iner al­lows you to cre­ate a me­mory po­ol for GPU me­mory al­lo­ca­ti­on. To avo­id me­mory al­lo­ca­ti­on and era­su­re du­ring com­pu­ting, Cha­iner pro­vi­des the abi­lity to use the CuPy me­mory po­ol as a stan­dard me­mory al­lo­ca­ti­on wit­ho­ut de­aling with me­mory al­lo­ca­ti­on. An in­tel­lec­tu­al system to de­ter­mi­ne the physi­cal and mec­ha­ni­cal pa­ra­me­ters of a mat­he­ma­ti­cal mo­del of non-isot­her­mal mo­is­tu­re transfer and vis­co­elas­tic de­for­ma­ti­on of ca­pil­lary-po­ro­us ma­te­ri­als was de­ve­lo­ped. It pro­vi­des the op­por­tu­nity to iden­tify pa­ra­me­ters of the ker­nels of cre­ep and re­la­xa­ti­on that is writ­ten as a li­ne­ar com­bi­na­ti­on of ex­po­nen­ti­al ope­ra­tors. The pro­po­sed al­go­rithm of appro­xi­ma­ti­on and ob­ta­ined cal­cu­la­ted ra­ti­os of rhe­olo­gi­cal be­ha­vi­or of wo­od by me­ans of mul­ti­la­yer neu­ral net­work with ex­po­nen­ti­al ac­ti­va­ti­on functi­ons in hid­den la­yers al­lows to incre­ase the ac­cu­racy of appro­xi­ma­ti­on of ex­pe­ri­men­tal cre­ep da­ta. The de­ve­lo­ped mat­he­ma­ti­cal mo­dels can be used to cre­ate an au­to­ma­ted systems of fi­ni­te-dif­fe­ren­ce cal­cu­la­ti­on of tem­pe­ra­tu­re and mo­is­tu­re con­tent, stress com­po­nents du­ring the drying of ca­pil­lary-po­ro­us ma­te­ri­als with ta­king in­to ac­co­unt the techno­lo­gi­cal pa­ra­me­ters of the drying agent.

[1]     Anag­nos­to­pou­los, I., Anag­nos­to­pou­los, C., & Lou­mos, V. (2004). Clas­sif­ying Web pa­ges emplo­ying a pro­ba­bi­lis­tic neu­ral net­work. IEEE Pro­ce­edings – Softwa­re, 151(3), (pp. 139–150).

[2]     Bod­yanskiy, Y. V., & Tyshchen­ko, O. K. (2019). A Hybrid Cas­ca­de Neu­ro-Fuzzy Net­work with Po­ols of Ex­ten­ded Neo-Fuzzy Neu­rons and its De­ep Le­ar­ning. In­ter­na­ti­onal Jo­ur­nal of Appli­ed Mat­he­ma­tics and Com­pu­ter Sci­en­ce, 29(3), 477–488. https://doi.org/10.2478/amcs-2019-0035

[3]     Chang, D.-J., Kan­tardzic, M. M., & Ou­yang, M. (2009). Hi­erarchi­cal Clus­te­ring with CU­DA/GPU. IS­CA PDCCS, 7–12.

[4]     Chap­man, B., Jost, G., & van der Pas, Ru­ud. (2008). Using OpenMP: por­tab­le sha­red me­mory pa­ral­lel prog­ram­ming (Sci­en­ti­fic and En­gi­ne­ering Com­pu­ta­ti­on), 2(3), 43–48. Cam­brid­ge, Mas­sac­hu­setts: The MIT Press.

[5]     Díaz, E., Bro­tons, V., & Tomás, R. (2018). Use of ar­ti­fi­ci­al neu­ral net­works to pre­dict 3-D elas­tic settle­ment of fo­un­da­ti­ons on so­ils with incli­ned bed­rock. So­ils and Fo­un­da­ti­ons, 58(6), 1414–1422. https://doi.org/10.1016/j.sandf.2018.08.001

[6]     Ger­bec, D., Gas­pe­ric, S., & Smon, I. (2005). Al­lo­ca­ti­on of the lo­ad pro­fi­les to con­su­mers using pro­ba­bi­lis­tic neu­ral net­works. IEEE Tran­sac­ti­ons on Po­wer Systems, 20(2), (pp. 548–555).

[7]     Gu, L., Li, X., & Si­egel, Ja. (2010). An em­pi­ri­cally tu­ned 2D and 3D FFT lib­rary on CU­DA GPU. Pro­ce­edings of the 24th ACM In­ter­na­ti­onal Con­fe­ren­ce on Su­per­com­pu­ting – ACM, Tsu­ku­ba, Ja­pan. – Ju­ne 01–04. New York, (pp. 305–314).

[8]     Hay­kin, S. (2013). Neu­ral Net­work a compre­hen­si­ve fo­un­da­ti­on (2nd ed.). Pren­ti­ce Hall, 426 p.

[9]     Hong, S. G., Kim, S. W., & Lee, J. J. (2015). The Mi­ni­mum Cost Path Fin­ding Al­go­rithm Using a Hop­fi­eld Type Neu­ral Net­work. Pro­ce­edings IEEE In­ter­na­ti­onal Con­fe­ren­ce on Fuzzy, Systems 4, (pp. 719–726).

[10]  Hu, Z., Bod­yanskiy, Y., & Tyshchen­ko, O. K. (2019). Self-le­ar­ning pro­ce­du­res for a ker­nel fuzzy clus­te­ring system. Ad­van­ces in In­tel­li­gent Systems and Com­pu­ting, 754, 487–497.

[11]  Spo­oner, J. T., Mag­gi­ore, M., Ordóñez, R., & Pas­si­no, K. M. (2002). Stab­le Adap­ti­ve Control and Es­ti­ma­ti­on for Non­li­ne­ar Systems: Neu­ral and Fuzzy Appro­xi­ma­tor Techniq­ues. John Wi­ley & Sons, Inc., 236 p.

[12]  Krste, A., et al. (2016). The Landsca­pe of Pa­ral­lel Com­pu­ting cRe­se­arch: A Vi­ew from Ber­ke­ley Uni­ver­sity of Ca­li­for­nia, Ber­ke­ley. Techni­cal cRe­port No. UCB/EECS-2016-183.

[13]  Na­bi­an, M. A., & Me­ida­ni, H. (2017). De­ep Le­ar­ning for Ac­ce­le­ra­ted Re­li­abi­lity Analysis of Infrastruc­tu­re Net­works. Com­pu­ter-Aided Ci­vil and Infrastruc­tu­re En­gi­ne­ering, 33(6), 443–458. https://doi.org/10.1111/mice.12359

[14]  Nu­ka­da, A., & Mat­suo­ka, S. (2009). Au­to-tu­ning 3-D FFT lib­rary for CU­DA GPUs. Pro­ce­edings of the Con­fe­ren­ce on High Per­for­man­ce Com­pu­ting Net­wor­king, Sto­ra­ge and Analysis – ACM, No­vem­ber 14–20, 2009, Portland, Ore­gon. – New York, (pp. 1–30).

[15]  Plas, D. V. (2018). Python Da­ta Sci­en­ce Handbo­ok: Es­sen­ti­al To­ols for Wor­king with Da­ta. St. Pe­tersburg: Pi­ter, 576 p.

[16]  Shymanskyi, V., & Protsyk, Yu. (2018). Si­mu­la­ti­on of the He­at Con­duc­ti­on Pro­cess in the Clay­di­te-Block Construc­ti­on with Ta­king In­to Ac­co­unt the Frac­tal Struc­tu­re of the Ma­te­ri­al. Com­pu­ter Sci­en­ce and In­for­ma­ti­on Techno­lo­gi­es: XIII-th In­ter­na­ti­onal Sci­en­ti­fic and Techni­cal Con­fe­ren­ce (CSIT-2018), Lviv Uk­ra­ine, (pp. 151–154).

[17]  So­ko­lovskyy, Y., Bo­retska, I., Gay­vas, B., Shymanskyi, V., & Gre­gus, M. (2019). Mat­he­ma­ti­cal mo­de­ling of he­at transfer in ani­sot­ro­pic bi­ophysi­cal ma­te­ri­als, ta­king in­to ac­co­unt the pha­se tran­si­ti­on bo­un­dary. CE­UR Workshop Pro­ce­edings 2488, (pp. 121–132).

[18]  So­ko­lovskyy, Ya., Mokrytska, O., & Krishta­po­vich, V. (2015). Mat­he­ma­ti­cal Si­mu­la­ti­on of De­for­ma­ti­on and Re­lax­ti­on pro­cess in ca­pil­larypo­ra­us ma­te­ri­als. Pro­ce­edings of the in­for­ma­ti­on Con­fe­ren­ce on Com­pu­ter Sci­en­ce and In­for­ma­ti­on Techno­lo­gi­es (CSIT 2015), Lviv Uk­ra­ine, (pp. 52–56).

[19]  So­ko­lovskyy, Ya., Nec­he­pu­ren­ko, A., & Zdolbytskyy, A. (2017). Softwa­re si­mu­la­ti­on of he­at mass transfer using pa­ral­lel com­pu­ting techno­lo­gi­es. Visnyk NULP: Komp'iu­ter­ni systemy. Te­oriia i praktyka, 923, 34–43.

[20]  So­ko­lovskyy, Ya., & Sin­kevych, O. (2018). Softwa­re and al­go­rithmic sup­port for rep­re­sen­ta­ti­on of CAD mo­dels in 2D von Neu­mann ne­ighbor­ho­od. CE­UR Workshop Pro­ce­edings 2300, (pp. 215–218).

[21]  So­ko­lowskyi, Ya., Shymanskyi, V., & Lev­kovych, M. (2016). Mat­he­ma­ti­cal mo­de­ling of non-iso­ter­mal mo­is­tu­re transfer and vis­co-elas­tic de­for­ma­ti­on in the ma­te­ri­als with frac­tal struc­tu­re. Com­pu­ter Sci­en­ce and In­for­ma­ti­on Techno­lo­gi­es. XI-Th In­ter­na­ti­onal Sci­en­ti­fic and Techni­cal Con­fe­ren­ce (CSIT-2016), Lviv, Uk­ra­ine, (pp. 91–95).

[22]  Ti­an, B., Azi­mi-Sad­ja­di, M. R., & Von­der Ha­ar, T. H. (2010). Tem­po­ral up­da­ting sche­me for pro­ba­bi­lis­tic neu­ral net­work with appli­ca­ti­on to sa­tel­li­te clo­ud clas­si­fi­ca­ti­on. IEEE Tran­sac­ti­ons on Neu­ral Net­works, 11(3–4), (pp. 903–920).

[23]  Tkac­hen­ko, R., Izo­nin, I., Kryvinska, N., Chop­yak, V., Lo­toshynska, N., & Danylyuk, D. (2018). Pi­ece­wi­se-li­ne­ar Appro­ach for Me­di­cal In­su­ran­ce Costs Pre­dic­ti­on using SGTM Neu­ral-Li­ke Struc­tu­re. In: Shak­hovska, N., Mon­te­neg­ro, S., Estève, Ya., Sub­bo­tin, S., Kryvinska, N., Izo­nin, I. (Eds.): In­for­ma­tics & Da­ta-Dri­ven Me­di­ci­ne. Pro­ce­edings of the 1st In­ter­na­ti­onal Workshop IDDM (IDDM 2018), Lviv, Uk­ra­ine, No­vem­ber 28–30, (pp. 170–179).

[24]  Tkac­hen­ko, R., & Izo­nin, I. (2019). Mo­del and Prin­cip­les for the Imple­men­ta­ti­on of Neu­ral-Li­ke Struc­tu­res ba­sed on Ge­omet­ric Da­ta Transfor­ma­ti­ons. In: Hu, Z. B., Pe­to­uk­hov, S., (Eds.): Ad­van­ces in Com­pu­ter Sci­en­ce for En­gi­ne­ering and Edu­ca­ti­on. Ad­van­ces in In­tel­li­gent Systems and Com­pu­ting. Sprin­ger, Cham. (ICCSEEA 2018). Vol. 754, (pp. 578–587).

[25]  Yang, C.-T., Hu­ang, C.-L., & Lin, C.-F. (2011). Hybrid CU­DA, OpenMP, and MPI pa­ral­lel prog­ram­ming on mul­ti­co­re GPU clus­ters. Com­pu­ter physics com­mu­ni­ca­ti­ons, 1, 266–269. https://doi.org/10.1016/j.cpc.2010.06.035

[26]  Zhi­yi, Y., Zhu, Y., & Pu, Y. (2008). Pa­ral­lel ima­ge pro­ces­sing ba­sed on CU­DA. Com­pu­ter Sci­en­ce and Softwa­re En­gi­ne­ering, 3, 198–201. https://doi.org/10.1109/CSSE.2008.144