Optimization of quasilinear stochastic control-nonlinear diffusion systems

详细

Consideration was given to the problem of optimal control of the quasilinear stochastic continuous-time systems whose coefficients in the general case depend nonlinearly on the program control. A special case of this problem is represented by optimization of the incompletefeedback control strategies. An optimization algorithm was proposed based on the method of gradient descent in the functional space. Its relaxation property was substantiated in theoretical terms. The necessary optimality conditions are formulated and proved within the framework of the formulated problem.