-
Regularity of Optimal Solutions and the Optimal Cost for Hybrid Dynamical Systems via Reachability Analysis
Authors:
Berk Altın,
Ricardo G. Sanfelice
Abstract:
For a general optimal control problem for dynamical systems with hybrid dynamics, we study the dependency of the optimal cost and of the value function on the initial conditions, parameters, and perturbations. We show that upper and lower semicontinuous dependence of solutions on initial conditions -- properties that are captured by outer and inner well-posedness, respectively -- lead to the exist…
▽ More
For a general optimal control problem for dynamical systems with hybrid dynamics, we study the dependency of the optimal cost and of the value function on the initial conditions, parameters, and perturbations. We show that upper and lower semicontinuous dependence of solutions on initial conditions -- properties that are captured by outer and inner well-posedness, respectively -- lead to the existence of a solution to the hybrid optimal control problem and upper/lower semicontinuity of the optimal cost. In particular, by exploiting properties of finite horizon reachable sets for hybrid systems, we show that the optimal cost varies upper semicontinuously when the hybrid system is (nominally) outer well-posed, and lower semicontinuously when it is (nominally) inner well-posed and an additional assumption requiring partial knowledge of solutions. Consequently, when the system is both (nominally) inner and outer well-posed and the aforementioned assumption holds, the optimal cost varies continuously, and optimal solutions vary upper semicontinuously. We further show that even in the absence of this solution-based assumption, the optimal cost can be continuously approximated. The results are demonstrated in finite horizon optimization problems with hybrid dynamics, both theoretically and numerically.
△ Less
Submitted 20 December, 2021;
originally announced December 2021.
-
Solutions and Reachable Sets of Hybrid Dynamical Systems: Semicontinuous Dependence on Initial Conditions, Time, and Perturbations
Authors:
Berk Altın,
Ricardo G. Sanfelice
Abstract:
The sequential compactness afforded hybrid systems under mild regularity constraints guarantee outer/upper semicontinuous dependence of solutions on initial conditions and perturbations. For reachable sets of hybrid systems, this property leads to upper semicontinuous dependence with respect to initial conditions, time, and perturbations. Motivated by these results, we define a counterpart to sequ…
▽ More
The sequential compactness afforded hybrid systems under mild regularity constraints guarantee outer/upper semicontinuous dependence of solutions on initial conditions and perturbations. For reachable sets of hybrid systems, this property leads to upper semicontinuous dependence with respect to initial conditions, time, and perturbations. Motivated by these results, we define a counterpart to sequential compactness and show that it leads to lower semicontinuous dependence of solutions on initial conditions and perturbations. In the sequel, it is shown that under appropriate assumptions, reachable sets of systems possessing this novel property depend lower semicontinuously on initial conditions, time, and perturbations. When those assumptions fail, continuous approximations of reachable sets turn out to be still possible. Necessary and sufficient conditions for the introduced property are given by a combination of geometric constraints, regularity assumptions, and tangentiality conditions. Further applications to simulations and optimal control of hybrid systems are discussed.
△ Less
Submitted 16 October, 2022; v1 submitted 26 March, 2021;
originally announced March 2021.
-
Exponential Stability of Nonlinear Differential Repetitive Processes with Applications to Iterative Learning Control
Authors:
Berk Altın,
Kira Barton
Abstract:
This paper studies exponential stability properties of a class of two-dimensional (2D) systems called differential repetitive processes (DRPs). Since a distinguishing feature of DRPs is that the problem domain is bounded in the "time" direction, the notion of stability to be evaluated does not require the nonlinear system defining a DRP to be stable in the typical sense. In particular, we study a…
▽ More
This paper studies exponential stability properties of a class of two-dimensional (2D) systems called differential repetitive processes (DRPs). Since a distinguishing feature of DRPs is that the problem domain is bounded in the "time" direction, the notion of stability to be evaluated does not require the nonlinear system defining a DRP to be stable in the typical sense. In particular, we study a notion of exponential stability along the discrete iteration dimension of the 2D dynamics, which requires the boundary data for the differential pass dynamics to converge to zero as the iterations evolve. Our main contribution is to show, under standard regularity assumptions, that exponential stability of a DRP is equivalent to that of its linearized dynamics. In turn, exponential stability of this linearization can be readily verified by a spectral radius condition. The application of this result to Picard iterations and iterative learning control (ILC) is discussed. Theoretical findings are supported by a numerical simulation of an ILC algorithm.
△ Less
Submitted 12 October, 2017; v1 submitted 19 January, 2017;
originally announced January 2017.
