Znachitel'noe chislo avariy obuslovleno otkazami unikal'nykh mekhanicheskikh sistem (UMS). Pod UMS ponimayutsya sistemy, izgotavlivaemye v 1 10 ekzemplyarakh, ekspluatiruemye v otlichayushchikhsya usloviyakh i realizuyushchie ekstremal'nye tekhnologicheskie operatsii. Predotvrashchenie avariy svyazano s zadachey ponizheniya veroyatnosti otkazov UMS, reshenie kotoroy v znachitel'noy stepeni zavisit ot tochnosti diagnostirovaniya i prognozirovaniya dinamiki ikh tekhnicheskogo sostoyaniya. Osobennost'yu zadachi yavlyaetsya otsutstvie statisticheskoy informatsii o dinamike izmeneniya tekhnicheskogo sostoyaniya, chto obuslovlivaet neobkhodimost' ispol'zovaniya metodov iskusstvennogo intellekta, v chastnosti,ekspertnykh sistem. Primenenie sovremennykh tekhnologiy imitatsionnogo modelirovaniya yavlyaetsya odnim iz sposobov diagnostirovaniya i prognozirovaniya tekhnicheskogo sostoyaniya UMS. Agentnyy podkhod k modelirovaniyu pozvolyaet issledovat' povedenie ob"ektov, soderzhashchikh bol'shoe chislo elementov i obladayushchikh individual'nym povedeniem. V rabote predstavlena intellektual'naya programmnaya sistema, realizuyushchaya agentnoe modelirovanie dinamiki tekhnicheskogo sostoyaniya na primere oborudovaniya neftekhimicheskikh proizvodstv.