Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Statistical results of parallel revevant backward inference

Keywords:

S.Yu. Bolotova – Ph. D. (Phys.-Math.), Lecturer University, Voronezh State University (VSU). E-mail: bolotova.svetlana@gmail.com


The relevant backward inference, based on the solution of logical equations, is an effective tool for research and optimization of production-logical systems. It can significantly reduce the number of slow-running queries to an external source of information. Previous papers describe the implementation of a multi-threaded algorithm for relevant backward inference. It is based on parallelizing steps in constructing a set of facts that are required in the inference. This article describes results of the statistical data research, obtained by performing a number of tests with using the relevant backward inference and its parallel implementation. It presents illustrations, charts and graphs, generated by data processing in the Statistica. The results for multi-threading implementation show a 20-25% average reduction of time required to find a solution. These figures are considered to be statistical indicators, as the classical backward inference allows to obtain the result immediately in some cases.
References:

 

  1. CHechkin A.V. Matematicheskaja informatika. M.: Fizmatlit. 1991. 416 s.
  2. Beniaminov E.M. Algebraicheskie metody v teorii baz dannykh i predstavlenii znanijj. M.: Nauchnyjj mir. 2003. 184 s.
  3. Bolotova S.JU.Algebraicheskaja model relevantnogo obratnogo vyvoda na osnove reshenija uravnenijj // Matematicheskoe modelirovanie. 2012. T. 24. № 12. S. 3–8.
  4. Bolotova S.JU., Makhortov S.D. Algoritmy relevantnogo obratnogo vyvoda, osnovannye na reshenii produkcionno-logicheskikh uravnenijj // Iskusstvennyjj intellekt i prinjatie reshenijj. 2011. № 2. S. 40–50.
  5. Forgy C.L. Rete: A fast algorithm for the many pattern/many object pattern match problem // Artificial Intelligence. 1982. № 19(1). P. 17–37.
  6. Miranker D. Performance estimates for the DADO machine: A comparison of TREAT and Rete // Fifth Generation Computer Systems, ICOT. Tokyo. 1984.
  7. Lee Y.H. Integration of Forward and Backward Inferences Using Extended Rete Networks // Industrial and Engineering Applications of Artificial Intelligence and Expert Systems, 9th, Proceedings of the Ninth International Conference. 1997. P. 339–345.
  8. Bolotova S.JU. Primenenie mnogopotochnosti v relevantnom LP-vyvode // Nejjrokompjutery. Razrabotka. primenenie. 2013. № 9. S. 53–57.
  9. Bolotova S.JU., Makhortov S.D. Parallelnye algoritmy relevantnogo LP-vyvoda // Programmnaja inzhenerija. 2014. № 7. S. 17–24.
  10. Ajjvazjan S.A.Prikladnaja statistika v 3-kh tomakh. Spravochnoe izdanie. M.: Finansyistatistika. 1989. 608 s.

 

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio