Differential inclusions and quasi-Lyapunov functions
Martin Ivanov, Mikhail Krastanov, Nadezhda Ribarska · 2025
Read Differential inclusions and quasi-Lyapunov functions by Martin Ivanov, Mikhail Krastanov, Nadezhda Ribarska (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-06100.txt.
Introduction
Martin Ivanov, Mikhail Krastanov, and Nadezhda Ribarska prove that a broad class of differential inclusions always has a solution. A differential inclusion generalizes an equation: the rate of change may be any of a set of allowed values. Their right-hand sides are uniformly bounded and take nonempty closed sets, which need not be convex.
Nonconvex values are the hard case, since many standard existence proofs lean on convexity. Their sufficient condition handles it, and a classic Olech-type result falls out as a corollary. To show the theorem earns its keep, they give an example rooted in the Fuller problem from optimal control theory. That problem is famous for its chattering optimal controls, making it a fitting stress test for inclusion theory.
Catalog record
| CREATOR | Martin Ivanov, Mikhail Krastanov, Nadezhda Ribarska |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-06100
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Differential inclusions and quasi-Lyapunov functions Martin Ivanov, Mikhail Krastanov, Nadezhda Ribarska Abstract A sufficient condition for existence of a solution of a differential inclusion with a uniformly bounded right-hand side that has nonempty closed (possibly nonconvex) values is obtained. A new way to deal with the so-called “bad points” is proposed. An Olech-type result is obtained as a corollary. An example, which originates from the Fuller problem from optimal control theory, is given to demonstrate the applicability of the main result. Keywords: differential inclusions with nonconvex right-hand side, existence of solutions, quasi-Lyapunov function, invariance, Fuller problem 2020 Mathematics Subject Classification: 34A36, 34A60 0 0 footnotetext: Martin Ivanov Faculty of Mathematics and Informatics, Sofia University “St. Kliment Ohridski”, 5 James Bourchier Blvd., 1164 Sofia, Bulgaria, e-mail: [email protected] Mikhail Krastanov Faculty of Mathematics and Informatics, Sofia University “St. Kliment Ohridski”, 5 James Bourchier Blvd., 1164 Sofia, Bulgaria and Institute of Mathematics and Informatics, Bulgarian Academy of Sciences, Acad. G. Bonchev str., bl. 8, 1113 Sofia, Bulgaria, e-mail: [email protected] Nadezhda Ribarska Faculty of Mathematics and Informatics, Sofia University “St. Kliment Ohridski”, 5 James Bourchier Blvd., 1164 Sofia, Bulgaria and Institute of Mathematics and Informatics, Bulgarian Academy of Sciences, Acad. G. Bonchev str., bl. 8, 1113 Sofia, Bulgaria, e-mail: [email protected] 1 Introduction We study the existence of local solutions of differential inclusions of the form \dot{x}\in F(x),\,x(0)=x_{0}, where x_{0}\in\mathbb{R}^{n} and F:\mathbb{R}^{n}\rightrightarrows\mathbb{R}^{n} is a multi-valued mapping with nonempty values, i.e. the existence of T>0 and an absolutely continuous function x:[0,T]\to\mathbb{R}^{n} such that x(0)=x_{0} and \dot{x}(t)\in F(x(t)) for almost every t\in[0,T] . It is well known that there is a solution for the case of upper semi-continuous right-hand side with nonempty convex compact values. The question whether solutions exist for the case of continuous right-hand side with nonconvex values was solved by Filippov in 1971 (cf. [ 10 ] ). This was generalized to the case of lower semi-continuous right-hand side by Bressan (1980, cf. [ 4 ] ) and by Lojasiewicz (1980, cf. [ 13 ] ) (more detailed historical remarks can be found in [ 9 ] ). A longstanding open problem is the existence of a solution of a differential inclusion with upper semi-continuous right-hand side with nonconvex values. As can be seen from the well-known Filippov’s examples (cf. [ 9 ] ), such a solution does not always exist unless some additional assumptions are made. Existence results for differential inclusions with mixed-type conditions on the right-hand side were proved by Olech (1975, cf. [ 15 ] ) and Lojasiewicz (1985, cf. [ 14 ] ) (see also Corollary 6.4 from [ 9 ] and the remarks after it). These results unify the upper semi-continuous and lower semi-continuous cases. For example, in Lojasiewicz’s result, lower semi-continuity on an open set and convexity and upper semi-continuity on its complement are assumed. In [ 3 ] Bivas generalized all known such theorems in the finite-dimensional case. In the present paper, we propose a theorem of this type, but instead of convexity, we assume that the field “moves away” from the set of points, where the right-hand side lacks lower semi-continuity. We make use of the formalism proposed in [ 11 ] , which in a way is a development of the patchy vector fields of Ancona and Bressan (cf. [ 1 ] ). This approach consists in constructing a “suitable” relatively open partition of the domain of the multi-valued mapping and approximating it on each element of the partition with an upper semi-continuous mapping which has convex compact values. This approximation mapping should be “not far” from the original one, and with respect to it the element of the partition should be weakly forward invariant (see Definition 2.7 below). This construction works well in the case where the relatively open partition is locally finite. In such a case, a uniform limit of \varepsilon -solutions is an \varepsilon -solution of the differential inclusion. For the case where locally finite partitions of the above kind cannot be constructed, a more complicated structure (consisting of countably many relatively open partitions, each refining the previous one, see Definition 2.8 below) is developed in [ 11 ] . In the present paper, we use the techniques from [ 11 ] to prove a sufficient condition for existence of a solution of a differential inclusion with uniformly bounded right-hand side that has nonempty closed values. We propose a different way of dealing with the “bad points”. In contrast to the assumption of convex values of the mapping on the set of “bad points” in [ 11 ] , we assume the existence of a function that keeps the \varepsilon -solutions of the inclusion starting from “good points” away from the set of “bad points”. In this way, we prove that it is possible to construct an \varepsilon -solution starting from a point on the boundary of the set of “bad points”. The function in question (called quasi-Lyapunov function) is assumed to be piecewise smooth. The reason is the fact that this research began as a search for conditions guaranteeing the existence of solutions of inclusions whose right-hand side belongs to some o-minimal structure (for example, subanalytic sets). In such a situation, piecewise smoothness is a natural assumption. It is possible to look for non-smooth variants. We show that the main result is applicable to an inclusion that originates from the optimal feedback for the Fuller problem from optimal control theory. This problem is one of the first examples of chattering control – the optimal control switches infinitely many times on a finite time interval (cf. [ 17 ] ). We construct a quasi-Lyapunov function in this example which is semi-algebraic. The structure of the paper is as follows. In the second section, the necessary preliminary definitions and results are stated. The third section contains the main result of the paper. A definition of a quasi-Lyapunov function is given. In the presence of such a function, we prove that there exists an \varepsilon -solution of the inclusion starting from a “bad point”. Under an additional assumption that involves successive refinements of the partition, the limit as \varepsilon\to 0 is taken and a solution of the inclusion is obtained. As a consequence, we obtain a result of Olech-type. The fourth section contains an application of the main result. The fifth section contains some concluding remarks. 2 Preliminaries The multi-valued mappings under consideration are defined on a nonempty subset D of \mathbb{R}^{n} and the solutions of the inclusions have to stay in D in order to be well defined. Moreover, one of the key ideas of the presented approach is the weak forward invariance of the elements of the partitions with respect to the corresponding approximating mappings, so we recall the following notions related to invariance. Definition 2.1 ( [ 11 ] ) . Let D\subseteq\mathbb{R}^{n} and \varphi:[0,T)\to\mathbb{R}^{n} be an absolutely continuous function. It is said that the set D is forward invariant with respect to \varphi if for each t\in[0,T) with \varphi(t)\in D there exists \delta>0 such that \varphi(\tau)\in D for all \tau\in[t,t+\delta) . Definition 2.2 ( [ 7 ] ) . Let S\subseteq D\subseteq\mathbb{R}^{n} and F:D\rightrightarrows\mathbb{R}^{n} be a multi-valued mapping. It is said that S is weakly forward invariant with respect to