Mathematical Fundamentals of Optimal Control

Abstract

To describe and study dynamic systems, the notions of system state, control effort and performance measure must be clarified. System state is a set of parameters that characterize the system at each time. The state parameters vary gradually and cannot instantly jump. The evolution of the state parameters, or state variables, as we will call them further, obeys dynamic laws following from the nature of the system considered as a moving or evolutionary object. For example, phase coordinates and momenta of mechanical systems obey Newton's laws or, in a more general case, Euler-Lagrange dynamic equations that underlie the motion of mechanical systems. In turn, manufacturing systems obey the laws of conservation, expansion or deterioration of mass. For example, the amount of product stored in a buffer is considered to be a state variable because its value changes in time in accordance with the mass conservation law. Indeed, the increment of product mass in the buffer at any time is equal to the sum of incoming and outgoing product flows through the buffer. For the same reasons we consider technological capabilities of machines and capacities of aggregate production as state variables which obey the laws of expansion and deterioration of mass. In modeling, these laws take on the form of dynamic equations with initial system states assumed to be known.