Differential Equations with Piecewise Constant Argument

• DEPCA is a hybrid dynamical system that fuses continuous differential equations with piecewise constant sampling, bridging ODEs and discrete maps.
• It employs reduction techniques to transform the problem into a discrete map framework, ensuring rigorous criteria for existence, uniqueness, periodicity, and oscillatory behavior.
• DEPCA methods underpin advanced numerical schemes and stability analyses, with applications ranging from control systems and population dynamics to neural networks.

A differential equation with piecewise constant argument (DEPCA) is a hybrid dynamical system in which the evolution depends both on the continuous state and on past or future states sampled at discrete, piecewise-constant grid points. DEPCA theory has matured into a rigorous discipline, with a wealth of results on existence, uniqueness, asymptotic behavior, periodicity, oscillation, stability, and numerical approximation. This article surveys the modern theory with an emphasis on structural definitions, solution frameworks, functional-analytic approaches, oscillatory and asymptotic properties, and advanced applications, referencing principal developments in the arXiv literature.

1. Mathematical Structure, Motivation, and Fundamental Definitions

DEPCA systems are distinguished by the inclusion of a discrete-time "memory" or sampled component in their evolution law. A canonical scalar case is

$$x'(t) = a\,x(t) + b\,x([t]) + f(t), \qquad t \in \mathbb{R}$$

where $[t]$ denotes the greatest integer less than or equal to $t$. More generally, one considers

$$x'(t) = A(t)x(t) + f(t, x([t])), \qquad t\in\mathbb{R}$$

with $A(t)$ continuous and $f$ satisfying appropriate regularity and Lipschitz conditions (Chavez et al., 2013).

For generalized arguments, the "sampling" map may be constructed using a sequence $\{t_n\}$ with

$$\theta(t) = t_n, \;\;\; \text{for } t \in [t_n, t_{n+1})$$

allowing for irregular or state-dependent grids (Castillo et al., 2014, Hartung, 4 Jan 2026). Models may be advanced, delayed, or mixed, with $\gamma(t)$ potentially given by

$\gamma(t) = \zeta_n, \;\;\; t \in [t_n, t_{n+1})$

where $\zeta_n \in [t_n, t_{n+1}]$ (Torres, 2024, Naranjo, 7 Apr 2025).

Motivations for DEPCA include

• Population dynamics with discrete census/sampling,
• Control systems with digital implementation,
• Hybrid or semi-discrete neural networks,
• Nonlocal or PT-symmetric quantum systems with reflection and sampled memory (Cabada et al., 20 Jan 2025, Cabada et al., 19 Jan 2026).

2. Core Analytical Techniques: Reduction to Discrete Maps and Solution Reconstruction

The defining feature of DEPCA—the piecewise constant argument—yields a hybrid structure which is exploited via reduction to difference equations. On each interval $[n, n+1)$ the system becomes an ODE, with the evolution controlled by constant sampled values. For linear and certain nonlinear cases, the state at integer nodes $x(n)$ evolves via a discrete map derived from matching at endpoints: $x(n+1) = C(n)x(n) + h(n)$ where $C(n)$ is a transition matrix or scalar, and $h(n)$ encodes the effect of $f$ (Chavez et al., 2013, Castillo et al., 2014). The continuous solution is then reconstructed by

$$x(t) = Z(t, [t])x([t]) + \int_{[t]}^{t} e^{A(t-s)}f(s)\,ds$$

where $Z(t, [t])$ is the fundamental or transition matrix on $[n, t)$ (Chavez et al., 2013).

This reduction enables proving:

• Existence and uniqueness for bounded, periodic, almost periodic, almost automorphic, or remotely almost periodic solutions under discrete analogs of exponential dichotomy or topological conjugacy (Chavez et al., 2013, Jaure et al., 31 Jan 2026, Pinto et al., 2015, Zou et al., 2016).
• Stability and asymptotics by transferring properties from discrete maps to the continuous-time problem (Castillo et al., 2014, Torres, 2024).

Core Reduction Step	Associated Discrete Structure	Continuous-Time Representation
Matching at nodes $t_n$	$x(n+1) = C(n)x(n) + h(n)$	$x(t)$ via $Z(t, [t])x([t]) + \int$
Jump/impulse at $t_k$	$x(n) = (1+c_n)x(n^-)$	Piecewise continuity + impulse condition
Nonlinearity in $x([t])$	Non-autonomous difference eq.	Composition via operator framework

3. Functional Spaces: Periodicity, Almost Periodicity, and Advanced Regularity Classes

The sampling inherent in DEPCA systems introduces discontinuous (but regulated) dependence, requiring extensions of classical function spaces:

• Almost Automorphic (AA) and Discontinuous Z–Almost Automorphic (ZAA) spaces characterize bounded solutions with (possibly discontinuous) argument dependence (Chavez et al., 2013).
• Bohr/Stepanov Almost Periodic/Asymptotically Periodic solutions and their generalizations are constructed for inhomogeneous or nonlinear systems, generally via contraction mapping or spectral methods (Dimbour et al., 2017, Akhmet et al., 2015, Castillo et al., 2014).
• Remotely Almost Periodic (RAP) classes provide a relaxed notion, capturing the “far-apart-almost-repeat” property necessary in the hybrid setting, with adapted discrete RAP for the sequence space (Jaure et al., 31 Jan 2026).
• In models with advanced, delayed, or mixed arguments, periodicity or almost periodicity of both the coefficient functions and the grid structure (e.g., equipotentially almost periodic step sequences) is essential (Castillo et al., 2014, Akhmet et al., 2015).

AA and RAP solution existence is tied to:

• Existence of a bi-almost automorphic (or bi-RAP) discrete Green’s function for the reduced map,
• Exponential dichotomy for the difference system,
• Lipschitz or contraction properties of the nonlinearity (Chavez et al., 2013, Jaure et al., 31 Jan 2026).

4. Oscillation, Stability, and Asymptotic Behavior

A central question in DEPCA theory is the oscillatory or stabilizing nature of solutions. The foundational approach is to transfer oscillation criteria from the induced difference equation to the full DEPCA, exploiting the equivalence of sign changes in the discrete trajectory to the continuous evolution (Naranjo et al., 17 Jul 2025, Naranjo, 7 Apr 2025, Torres, 2024).

Sharp oscillation and nonoscillation criteria include:

• Erbe–Zhang type: $\liminf Q_n^* > \frac{k^k}{(k+1)^{k+1}}$ for delayed, Ladas–Philos–Sficas and Győri–Ladas criteria for advanced, and Wiener–Aftabizadeh-type for arbitrary mesh (Naranjo et al., 17 Jul 2025, Torres, 2024, Naranjo, 7 Apr 2025).
• Exponential dichotomy, Lyapunov exponents, and Floquet multipliers are computable from the discrete map and drive asymptotic classification (exponential decay, boundedness, instability) (Torres, 2024).
• Second-order and functional-analytic extensions yield oscillation conditions mirroring Leighton–Wintner results, with appropriate integral constraints on weight and comparison functions (Naranjo, 7 Apr 2025).

In non-autonomous, impulsive, or mixed-advance systems, hybrid criteria combine sign patterns, integrals on retarded/advanced subintervals, and properties of impulses.

5. Existence, Boundedness, and Topological Linearization of Solutions

The general solution theory encompasses:

• Linear systems: explicit variation-of-parameters formulas apply, incorporating the effect of the piecewise constant argument at each interval, and capturing both regular and impulse-driven DEPCAG (Torres et al., 2024, Castillo et al., 2014).
• Nonlinear systems: Banach-fixed-point methods and small-gain arguments (including in partially unbounded settings) demonstrate existence and stability of bounded and almost periodic solutions under contractive or spectral gap conditions (Zou et al., 2016, Pinto et al., 2015, Castillo et al., 2014).
• Topological conjugacies: Under suitable exponential dichotomy, generalizations of the Grobman–Hartman theorem yield strong topological and even Hölder equivalence between nonlinear DEPCA and their linear counterparts (Pinto et al., 2015, Zou et al., 2016). The conjugacy—constructed as a globally continuous homeomorphism—transfers stability and equivalence of attractor/repellor structures.
• Homoclinic and heteroclinic orbits: The existence of such objects is demonstrated via Banach-space contraction, with continuous-time behaviors mirroring the discrete orbit structure of an associated map (e.g., the logistic map) (Fen et al., 11 Mar 2025).

6. Numerical Analysis and Approximation Schemes

Numerical analysis of DEPCA centers on both qualitative and quantitative correspondence between continuous and discrete representations:

• The piecewise-constant-argument approach yields natural stepwise schemes matching the discrete difference equations, offering uniform first-order convergence when regularity conditions hold (Sepúlveda, 2016, Hartung, 4 Jan 2026).
• For Hamiltonian or energy-conserving structures, high-order, globally energy-preserving Runge-Kutta-type methods (Hamiltonian Boundary Value Methods, HBVMs) are constructed by Legendre polynomial expansions and Fourier truncation, controlling the propagation of discontinuities in the sampled argument (Gurioli et al., 2024).
• Error bounds are established (e.g., $O(h^{2s})$ for HBVM($k,s$)), and stability transfer from continuous to discrete is rigorously analyzed via Halanay-type integral inequalities (Sepúlveda, 2016, Gurioli et al., 2024).
• The EPCA approach demonstrates that solutions of the discretized “piecewise-argument” system converge uniformly to those of the original state-dependent delay/impulsive DDEs under monotonicity and Lipschitz assumptions (Hartung, 4 Jan 2026).

7. Advanced Models and Applications: Reflections, Nonlocality, and Generalized Structures

Recent research has extended DEPCA to settings that incorporate further structural nonlocality:

• Equations with involution or reflection combine $v(-t)$ and $v([t])$ under periodic conditions, requiring construction of multidimensional Green’s functions and rigorous classification of positivity/negativity regions—a critical property for applying fixed-point theorems for nonlinear problem existence (Cabada et al., 20 Jan 2025, Cabada et al., 19 Jan 2026).
• The constant-sign region of the Green’s function is demarcated explicitly as a function of parameters $(m, M, T)$, with implications for solution existence and uniqueness, and connections to Dirichlet eigenvalues in perturbed Schrödinger problems (Cabada et al., 19 Jan 2026).
• Models with adaptive, state-dependent delays and impulses are shown to admit robust existence, uniqueness, and convergence theory when approximated by EPCA-type schemes (Hartung, 4 Jan 2026).
• Multilevel almost periodic and Bohr–Stepanov regularity, as well as generalized arguments with arbitrary mesh or state-dependent grids, are handled by employing advanced function space and spectral techniques (Castillo et al., 2014, Jaure et al., 31 Jan 2026).

Conclusion and Future Perspectives

DEPCA theory unifies continuous and discrete dynamical phenomena, supporting an extensive theory based on reduction to discrete-time maps, functional-analytic solution concepts, and direct analytical criteria for oscillatory, periodic, and asymptotic behaviors. Open directions include the extension of the spectral and stability theory to pseudo–almost automorphic and stochastic settings, inclusion of neutral and state-dependent arguments, further development of topological and smooth conjugacy results, and refinement of numerical methods to preserve qualitative features under discretization (Chavez et al., 2013, Castillo et al., 2014, Gurioli et al., 2024).

Open problems remain concerning the complete spectral characterization of almost periodicity in the DEPCA context, the treatment of stochastic perturbations, hybrid random systems, and deeper links between hybrid, impulsive, and delay differential systems. Extension to infinite-dimensional settings and the rigorous analysis of more complex neural and quantum models with sampled, reflected, or nonlocal arguments are particularly active research fronts.