⚡ electrical engineering

Stability and Bifurcations of Planar Switched Linear and Homogeneous Systems

This paper establishes new explicit necessary and sufficient conditions for the uniform asymptotic stability of planar switched homogeneous systems under arbitrary switching, which are then leveraged to analyze codimension-one bifurcations and derive novel local and global stability results, including an analogue of Lyapunov's indirect method and a basin of attraction criterion, for specific classes of switched nonlinear systems.

Original authors: Ivan O. Shevchenko, Xinzhi Liu

Published 2026-07-15

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Ivan O. Shevchenko, Xinzhi Liu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are the conductor of a chaotic orchestra where the musicians don't just play notes; they suddenly swap instruments and sheet music mid-song. In the world of math and physics, this is called a switched system. Instead of one steady rule governing how a system moves, you have a collection of different rules (subsystems) that take turns controlling the action. The big question is: Will this system eventually calm down and settle at zero, or will it spiral out of control?

For decades, mathematicians have tried to answer this for flat, two-dimensional systems (like a ball rolling on a table). The new paper by Ivan O. Shevchenko and Xinzhi Liu acts like a master detective, finally solving the case with a set of crystal-clear, "yes-or-no" rules.

The Detective's Toolkit: The "Worst-Case" Scenario

Most previous attempts to solve this were like trying to predict the weather by looking at the average temperature. They gave "good enough" hints, but they couldn't tell you for sure if a storm was coming. This paper takes a different approach: it asks, "What is the absolute worst way this system could behave?"

Imagine the system is a hiker trying to get back to a campfire (the origin, or zero). The hiker has a map with different terrains (the subsystems). Some terrains pull the hiker toward the fire; others might push them away. The "worst-case" analysis imagines a mischievous switcher who changes the terrain at the exact moment the hiker is most vulnerable, trying to push them as far away from the fire as possible.

The authors prove that if the system can survive this mischievous switcher's best efforts, it is uniformly asymptotically stable. This means no matter how the rules are switched, the system will eventually return to zero.

The Two Golden Rules

The paper provides two specific conditions that must both be true for the system to be safe. Think of these as the two legs of a stool; if one breaks, the whole thing falls.

The "No-Go" Zones (Condition 1):
Imagine drawing lines on the ground. The paper identifies specific rays (lines shooting out from the center) where the different terrains might disagree on which way to push. If the system tries to switch between two specific terrains while standing on one of these lines, and the math says they are pushing in a "bad" direction (specifically, if a certain calculation called $\Delta_{p,q}(x)$ is not positive), the system is doomed. The authors show that for stability, these "bad" switches must never happen in the dangerous zones.
The "Shrinking" Loop (Condition 2):
Imagine the hiker is forced to run in a circle, switching terrains every time they cross a line. The paper calculates a "weight" for each leg of the journey: how much bigger or smaller the hiker gets after crossing from one line to the next.
- If you multiply all these "growth factors" together for a full circle, the result must be less than 1.
- If the product is greater than 1, the hiker grows bigger with every lap and never returns to the fire.
- If the product is exactly 1, the hiker runs in a perfect, endless loop (a periodic solution) and never settles down.

What the Paper Rules Out

The authors are very careful about what they don't claim. They explicitly state that their method is fundamentally limited to two dimensions (flat surfaces). You cannot use these specific rules to predict the behavior of a system in 3D space (like a drone flying in a room) or higher dimensions. The math gets too messy and the "rays" don't organize themselves as neatly in higher dimensions.

They also argue against the idea that you can always find a single "common Lyapunov function" (a universal energy score) for every switched system. While such functions exist for stable systems, finding them is often impossible to do in practice. Instead, this paper offers a direct, algorithmic way to check stability without needing to find that elusive universal score first.

The "Bifurcation" Moment: When Stability Breaks

One of the most exciting parts of the paper is how it explains how stability breaks. This is called a bifurcation.

Imagine you are slowly turning a dial (a parameter) on your system. As long as the dial is in the "safe" zone, the system is stable. But what happens exactly when you turn it just a tiny bit too far?

The Paper's Discovery: The system doesn't just slowly drift away. Instead, it suddenly snaps into a periodic orbit. It starts running in a perfect, endless circle right near the center.
The authors prove that if the "shrinking" rule (Condition 2) fails, or if the "No-Go" rule (Condition 1) fails, a stable system will immediately transform into a system that loops forever. It's like a spinning top that, instead of wobbling and falling, suddenly locks into a perfect, unending spin.

From Straight Lines to Curves (Nonlinear Systems)

The paper doesn't stop at simple, straight-line systems (linear). It uses these rules as a "local" test for more complex, curvy systems (nonlinear).

The Analogy: Imagine a rollercoaster that looks like a straight track for a tiny bit at the very top. If the straight track would be stable, the rollercoaster is safe right there at the top. If the straight track would be unstable, the rollercoaster will crash right there at the top.
The authors prove that for systems that look like straight lines when you zoom in close enough, these two golden rules are the ultimate test. If the linear version is stable, the complex version is locally stable. If the linear version fails, the complex version will have a loop or a crash nearby.

How Sure Are We?

The authors are extremely confident in their main findings. They haven't just simulated this on a computer or suggested it might work; they have mathematically proved it.

They provide necessary and sufficient conditions. This is the gold standard in math. It means:
- If the conditions are met, the system is stable (100% sure).
- If the conditions are not met, the system is not stable (100% sure).
- There are no "maybe" zones in their main theorem.

However, they are honest about the limits. They admit that while they can prove stability for 2D systems, they cannot yet extend this specific "worst-case" ray analysis to 3D or higher dimensions. They leave that as a puzzle for future mathematicians.

The Takeaway

This paper is like a new, super-accurate map for a specific type of terrain. It tells you exactly where the cliffs are and exactly how to avoid them. It replaces vague guesses with a clear, step-by-step checklist. If you have a flat, two-dimensional system that switches between different rules, you can now use these two simple rules to know with absolute certainty whether it will settle down or spin out of control forever. And if it does spin out, you know exactly that it will turn into a perfect, endless loop.

Problem Statement
The paper addresses the stability analysis of planar (two-dimensional) switched dynamical systems, specifically focusing on switched linear systems and switched homogeneous systems with a finite number of subsystems. While extensive literature exists on providing sufficient conditions for stability (often easier to verify but not exhaustive) or necessary and sufficient conditions that are often intractable or lack qualitative insight, there is a gap in understanding the precise mechanisms by which uniform asymptotic stability is lost. Specifically, the authors note that existing necessary and sufficient conditions for two-dimensional systems (e.g., Filippov, Pyatnitsky, Rapoport) either fail to provide insight into trajectory behavior during bifurcations or rely on assumptions that restrict the class of systems or require solving difficult nonlinear equations. Furthermore, little progress has been made in classifying codimension-one bifurcations of stability for switched systems with arbitrary switching signals.

Methodology
The authors employ a "worst-case switching analysis" approach, which is explicit and geometric in nature. The core methodology involves:

Geometric Decomposition: The state space $\mathbb{R}^2$ is partitioned based on a finite set of rays, denoted $S = \{S[1], \dots, S[n]\}$ . These rays are derived from the eigenvectors of the subsystem matrices, the zero sets of quadratic forms associated with the symmetric parts of the matrices, and the zero sets of determinants involving pairs of subsystem matrices.
Graph-Theoretic Representation: A weighted directed graph is constructed where vertices represent the sectors between these rays, and edges represent possible transitions between sectors under specific subsystems. Weights on the edges quantify the expansion or contraction of the trajectory norm as it crosses from one ray to another.
Auxiliary System Construction: To handle cases where certain transitions are impossible (i.e., weights do not exist) in the original system, the authors introduce an "auxiliary" switched system by appending subsystems with dynamics close to $-I_2$ (specifically matrices $R_{\delta+}$ and $R_{\delta-}$ ). This ensures that all necessary crossing times exist, allowing the application of their main stability theorem, after which they prove the stability properties of the original and auxiliary systems are equivalent.
Extension to Nonlinear Systems: The linear results are extended to $C^1$ switched nonlinear systems and homogeneous systems by utilizing converse Lyapunov theorems and analyzing the linearization at the origin or the asymptotic behavior at infinity.

Key Contributions and Results
The paper establishes several new theoretical results:

Necessary and Sufficient Conditions (Theorem 3.5): The authors prove a set of necessary and sufficient conditions for the uniform asymptotic stability of planar switched linear systems under arbitrary switching. These conditions are:
1. A geometric condition involving the determinant of pairs of subsystem matrices ( $\Delta_{p,q}(x) > 0$ ) within specific cones.
2. A product condition on the weights of cycles in the associated graph (the product of weights for any cycle must be less than 1).
Bifurcation Analysis (Theorem 4.1): The explicit nature of the derived conditions allows for a detailed classification of codimension-one bifurcations. The paper characterizes exactly how stability is lost: either through the violation of the determinant condition (leading to a one-dimensional subspace where equality holds) or the violation of the weight product condition. Crucially, the authors prove that if stability is lost due to these conditions failing with strict inequality, a periodic solution of the switched system must exist arbitrarily close to the origin.
Analogue of Lyapunov's Indirect Method (Theorem 4.2): For $C^1$ switched nonlinear systems with a common equilibrium at the origin, the paper formulates a test analogous to Lyapunov's indirect method. If the linearized switched system satisfies the stability conditions with strict inequality, the nonlinear system is locally uniformly asymptotically stable. If the conditions fail with strict inequality, the nonlinear system possesses a non-convergent (periodic) solution near the origin.
Global Stability and Basin of Attraction (Theorem 4.3 & 4.4): The results are extended to switched systems where subsystems are asymptotically linear at infinity, providing conditions for the existence of a bounded basin of attraction. Additionally, for homogeneous systems of degree $\beta > 1$ , the stability of the nonlinear system is shown to be equivalent to the stability of its linearization at the origin.

Significance and Claims
The authors claim that the primary novelty of their work lies in the explicit nature of their approach. Unlike previous methods that provide abstract conditions or rely on high-level existence proofs, this framework allows for:

Qualitative Insight: A clear visualization of how trajectories behave and how stability is lost, enabling the classification of bifurcations.
Algorithmic Verification: The conditions are stated in a way that can be checked computationally (e.g., computing eigenvalues, determinants, and weight products).
Generalization: The method unifies the study of linear and homogeneous systems and extends to specific classes of nonlinear systems without requiring the construction of a common Lyapunov function explicitly, though it relies on the existence of one.

The paper concludes that while the approach is fundamentally limited to two dimensions (due to the reliance on planar geometry and the specific structure of the ray set $S$ ), it provides a complete and transparent solution to the stability and bifurcation problem for planar switched systems, filling a gap left by previous works that either lacked bifurcation insight or imposed restrictive assumptions.