Use of mathematical models for accelerator extreme control systems

Problems of selecting models to construct algorithms of extremal control of charged-particle accelerators are discussed. Basing on the experience of the development of the extremal control of the ITEF synchrotron and the IFVE booster, deterministic, statistical, adaptive and simulative models are considered. The considered models make it possible to solve the following problems: selection and ranking of control systems, variation of which provides optimization of the output parameter and its connection with the control vector; the synthesis of the control algorithm and the integrated check-out of extremal control operating programs