objective function at a solution, while only depending on the parameters of the problem.[1][2] In a controlleddynamical system, the value function represents the optimal payoff of the system over the interval [t, t1] when started at the time-tstate variablex(t)=x.[3] If the objective function represents some cost that is to be minimized, the value function can be interpreted as the cost to finish the optimal program, and is thus referred to as "cost-to-go function."[4][5] In an economic context, where the objective function usually represents utility, the value function is conceptually equivalent to the indirect utility function.[6][7]
In a problem of
supremum
of the objective function taken over the set of admissible controls. Given , a typical optimal control problem is to
subject to
with initial state variable .[8] The objective function is to be maximized over all admissible controls , where is a Lebesgue measurable function from to some prescribed arbitrary set in . The value function is then defined as
with , where is the "scrap value". If the optimal pair of control and state trajectories is , then . The function that gives the optimal control based on the current state is called a feedback control policy,[4] or simply a policy function.[9]
Given this definition, we further have , and after differentiating both sides of the HJB equation with respect to ,
which after replacing the appropriate terms recovers the costate equation
where is
Newton notation for the derivative with respect to time.[12]
The value function is the unique viscosity solution to the Hamilton–Jacobi–Bellman equation.[13] In an online closed-loop approximate optimal control, the value function is also a Lyapunov function that establishes global asymptotic stability of the closed-loop system.[14]
Clarke, Frank H.; Loewen, Philip D. (1986). "The Value Function in Optimal Control: Sensitivity, Controllability, and Time-Optimality". SIAM Journal on Control and Optimization. 24 (2): 243–263.