Filchenko D. Modeling and Identification in Linear Quadratic simulation, forecasting and optimization problems

The new identification approaches for LQ simulation, forecasting and optimization problems for weak formalized systems have been developed in the given thesis. The conditions of matching linear time-invariant Cauchy problem with the decomposition of its solution have been found. The mixed integral and collocation identification models for LQ problems have been elaborated in terms of searching for efficient initial approximation and spectrum multicriterion regulation and have been extended over the class of problems with unknown initial conditions and control. The conditions of matching optimal control LQ problem for an optimization time range with identification problem of that model for a base time range have been found.