Optimal control of finite horizon type for a multidimensional delayed switching system (Q558580)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Optimal control of finite horizon type for a multidimensional delayed switching system
scientific article

    Statements

    Optimal control of finite horizon type for a multidimensional delayed switching system (English)
    0 references
    0 references
    1 July 2005
    0 references
    The author studies the control system governed by the ordinary differential system \[ y'(t) = f(y(t), z(t), \alpha(t)) \;(t \geq 0)\, , \quad y(0) = x, \] where \(y(t)\) is an \(n\)-dimensional vector function and \(\alpha(t)\) is a scalar control. The switchings of the scalar function \(z(t)\) depend on the trajectory \(y(t)\) through the law \(z(t) = h[y(t) \cdot S, w](t)\), with \(S\) an \(n\)-dimensional vector and \(w\) an initial condition for \(z(t)\). The functional to be minimized is \[ J(x, w, t, \alpha) = \int_0^t e^{- \lambda s}\ell(y(s), z(s), \alpha(s))\,ds + e^{- \lambda t} g(y(t), z(t)) \] over all trajectories with given initial condition. Several definitions of the value function are possible, bot none is in general continuous. The author shows that the value function of a suitable relaxed version of the problem is the unique lower semicontinuous viscosity solution of two coupled Hamilton-Jacobi equations.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    delayed switchings
    0 references
    optimal control
    0 references
    finite horizon
    0 references
    hysteresis
    0 references
    exit time
    0 references
    discontinuous viscosity solutions
    0 references
    0 references