Computer Science > Systems and Control
[Submitted on 28 May 2019 (this version), latest version 22 May 2020 (v2)]
Title:When do Trajectories have Bounded Sensitivity to Cumulative Perturbations
View PDFAbstract:A dynamical system is said to have bounded sensitivity if an additive disturbance leads to a change in the state trajectory that is bounded by a constant multiple of the size of the cumulative disturbance. It was shown in Sharifnassab, et al. (2019) that the (negative) (sub)gradient field of a piecewise constant and convex function with finitely many pieces has always bounded sensitivity. In this paper, we investigate sensitivity of some generalizations of this class of systems, as well as other classes of practical interest. In particular, we show that the result in Sharifnassab, et al. (2019) does not generalize to the (negative) gradient field of a convex function. More concretely, we provide examples of gradient field of a twice continuously differentiable convex function, and subgradient field of a piecewise constant convex function with infinitely many pieces, that have unbounded sensitivity. Moreover, we give a necessary and sufficient condition for a linear dynamical system to have bounded sensitivity, in terms of the spectrum of the underlying matrix. The proofs and the development of our examples involve some intermediate results concerning transformations of dynamical systems, which are also of independent interest. More specifically, we study some transformations that when applied on a dynamical system with bounded sensitivity, preserve the bounded sensitivity property. These transformations include discretization of time and spreading of a system; a transformation that captures approximate solutions of the original system in a certain sense.
Submission history
From: Arsalan Sharifnassab [view email][v1] Tue, 28 May 2019 11:24:03 UTC (223 KB)
[v2] Fri, 22 May 2020 04:32:01 UTC (324 KB)
Current browse context:
eess.SY
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.