Synthesis of an optimal control for continuous Markov processes by semidefinite quality criteria (Q1778321)

Optimization of control of continuous Markov processes by semidefinite quality functionals with nonadditive control cost functions is investigated. An equation for the loss function, its parameter for different integrands in the quality criterion, and its interrelation with the Fokker-Planck-Kolmogorov equation for the auxiliary noncontrollable Markov processes are determined. Examples are given. The presented control synthesis approach is effective because the nonlinear differential equation for the loss function is reduced to a linear parabolic partial differential equation.