A Bliss-type multiplier rule for constrained variational problems with time delay (Q1363601)

The optimization problem for delay systems: \[ \min J(y)= \int^{t_2}_{t_1} f(t, y(t), y'(t), y(t-\tau), y'(t-\tau)) dt \] such that \[ \begin{aligned} \phi_\beta(t, y(t), y'(t), y(t-\tau), y'(t-\tau)) & = 0,\;\beta= 1,\dots, m< n,\\ \phi_\mu(t_1, y(t_1), t_2,y(t_2)) & = 0,\;\mu= 1,\dots, p\leq 2n,\\ y(t) & = \alpha(t),\;t\in [t_1-\tau, t_1],\end{aligned} \] is considered and new necessary conditions are obtained in the form of a multiplier rule: a linear combination of \(f\) and \(\phi_1,\dots,\phi_m\) satisfy generalized Euler equations and appropriate boundary conditions. It is shown that every minimizing arc of the above problem satisfies a multiplier rule. The example \[ \min J(y)= {1\over 2} \int^2_0 y^2_2(t)dt \] such that \[ \begin{aligned} y_1'(t)+ y_1(t- 1)- y_2(t) & = 0,\\ y_1(t) & = 1,\;t\in [-1,0],\end{aligned} \] is solved.