A.N. Shmarin - Post-Graduate Student, Voronezh State University. E-mail: tim-shr@mail.ru S.D. Makhortov - Dr. Sc. (Phys.-Math.), Associate Professor, Head of Department, Voronezh State University (VSU). E-mail: sd@expert.vrn.ru

With increasing volumes and complexity of processed information, the use of artificial intelligence systems is becoming increa-singly important. At the same time one of the most common models of knowledge representation is productional. Algorithms for finding solutions in such systems are based on the inference engine. In practice, there are many problems for which obtaining in-formation about the state of the subject area is the resource-intensive operation. For example, these might be a request to the database, analysis of medical data or a complex sequence of computations. Often in similar tasks the solution (or the set of poss-ible solutions) is looked for on the basis of only part of the information about the state of the subject area. In view of this cir-cumstance it is required to optimally use limited resources. There are tasks of control complex objects and random processes (risk management, socio-economic systems), in which the control is based on the logical-probabilistic model of knowledge representation. For them, the result of the controlling influence can be predicted only with some probability. In such situations, it is necessary to maintain or lead system into the predetermined state, applying as little as possible controlling influences. The related problem is to minimize the number of requests for obtaining information about the state of the subject area during logical inference. This problem is NP-hard. For achievement of global mi-nimization there is the general method of LP-inference, which, unfortunately, has exponential computational complexity relative to the number of atomic facts in the knowledge base. However, its some modifications have polynomial complexity. The idea of cluster-relevant LP-inference is based on the calculation of specific estimates (indicators of relevance) for the limited subset of productions and the generalization of the obtained estimates on the entire set of productions. This paper introduces heuristic estimates allowing to take into account the structural relationships in the set of productions, and thus improve the degree of minimization of the number of requests that are executed in the process of logical inference.

 

1. Sowyer B., Foster D. Programming Expert Systems in Pascal // John Wiley & Sons, Inc. 1986. 186 p.
2. Makhortov S.D.,SHmarin A.N. Nechetkijj LP-vyvod i ego programmnaja realizacija // Programmnaja inzhenerija. 2013. № 12. S. 34-38.
3. Makhortov S.D. Osnovannyjj na reshetkakh podkhod k issledovaniju i optimizacii mnozhestva pravil uslovnojj sistemy perepisyvanija termov // Intellektualnye sistemy. 2009. T. 13. Vyp 1-4. S. 51-68.
4. Makhortov S.D. LP-struktury dlja obosnovanija i avtomatizacii refaktoringa v obektno-orientirovannom programmirovanii // Programmnaja inzhenerija. 2010. № 2. S. 15-21.
5. 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.
6. Makhortov S. D. Integrirovannaja sreda logicheskogo programmirovanija LPExpert // Informacionnye tekhnologii. 2009. № 12. C. 65-66.
7. Salijj V.N., Bogomolov V.A. Algebraicheskie osnovy teorii diskretnykh sistem. M.: Fizmatlit. 1997. 368 s.
8. Levi G., Sirovich F. Generalized And/Or Graphs. Artificial Intelligence Journal.1976. V. 7. P. 243-259.
9. KHarari F. Teorija grafov: Per. s angl. M.: Editorial URSS. 2003. 296 s.
10. Skiena S. Algoritmy. Rukovodstvo po razrabotke. Izd. 2-e. SPb.: BKHV-Peterburg. 2011. 720 s.
11. CHechkin A.V. Nejjrokompjuternaja paradigma informatiki // Nejjrokompjutery: razrabotka, primenenie. 2011. №7. S. 3-9.