Optimization of a fractional differential equation (Q2419407)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Optimization of a fractional differential equation
scientific article

    Statements

    Optimization of a fractional differential equation (English)
    0 references
    0 references
    0 references
    13 June 2019
    0 references
    The control system is \[ D^\gamma u(t) + \mathcal{A} \, u(t) = f(t) + z(t) \quad 0 \le t \le T \, , \qquad u(0) = \psi \] where $\mathcal{A}$ is a symmetric positive definite $n \times n$ matrix and $z(t)$ is the control. The fractional derivative is defined by \[ D^\gamma v(t) =\frac{1}{\Gamma}(1 - \gamma) \int_0^tv'(\zeta) d\frac{\zeta}{(t - \zeta)^\gamma} \] for $0 < \gamma < 1.$ The control problem is that of minimizing a quadratic functional \[ J(u, z) = \int_0^T \big( \|\mathcal{C} u(t) - u_d\|^2 + \mu \|z(t)\|^2 \big) dt \] $(\mathcal{C}$ a $m \times n$ matrix) subject to the control constraints $a \preceq z(t) \preceq b$, where $\lq\lq \preceq ``$ means $\lq\lq \le $ for all coordinates''. The authors study existence, uniqueness and regularity of the solutions of the system and set up a discretization scheme to solve the control problem. They show stability and consistence of the scheme and examine rates of convergence. For the entire collection see [Zbl 1402.49004].
    0 references
    linear quadratic optimization problem
    0 references
    fractional ordinary differential equation
    0 references
    control constraints. discretization
    0 references
    stability
    0 references
    consistence
    0 references

    Identifiers