Asymptotic controllability and optimal control (Q1943474)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Asymptotic controllability and optimal control
scientific article

    Statements

    Asymptotic controllability and optimal control (English)
    0 references
    0 references
    0 references
    20 March 2013
    0 references
    A control problem with a Lagrange functional \[ J(z,\alpha)= \int^T_0 l(z(t), \alpha(t))\,dt \] and a state equation \(\dot z(t)= f(z(t), \alpha(t))\), \(z(0)= x\) is studied. The controls \(\alpha(\cdot)\) are asssumed to be Borel-measurable with values in a compact control set \(A\). Additionally, a closed target set \(C\) is given. Admissible states have to fulfill the approximate final condition \(\lim_{t\to T^-}d(z(t), C)= 0\). Denote \(V(x)= \text{inf\,} J(x,\alpha)\). A crucial assumption is the sign condition \(l(z,\alpha)\geq 0\) for all \(z\), \(\alpha\). Through an inequality involving a positive number \(p\) and a minimum restraint function \(U(x)\) a condition for the asymptotic controllability is provided. Then, the value function is bounded by \(U/p\). The result has consequences for the corresponding Hamilton-Jacobi equation. Three examples illustrate the results.
    0 references
    optimal control
    0 references
    asymptotic controllability
    0 references
    Lyapunov function
    0 references
    value function
    0 references
    exit-time problems
    0 references
    Hamilton-Jacobi equation
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references