Mackey-Glass Benchmark Dynamics

• Mackey-Glass benchmark is a delay differential model featuring time-delay and nonlinear feedback that underpins studies of bifurcations, chaos, and multistability.
• Electronic, physical, and numerical implementations validate its prediction of fixed points, periodic oscillations, and chaotic regimes in complex systems.
• Control strategies such as proportional feedback and Pyragas control are applied to harness and modulate chaos for applications in secure communication and neuromorphic computing.

The Mackey-Glass benchmark refers to the class of delay differential equations (DDEs) introduced by Mackey and Glass to model physiological systems where the rate of change at time $t$ depends nonlinearly on the system's state at a previous time $t - \tau$. Since its inception, the Mackey-Glass equation has become an archetype for studying infinite-dimensional nonlinear dynamics, bifurcation theory, multistability, chaos, and stochastic phenomena in time-delayed feedback systems. Numerous electronic, physical, and numerical implementations have solidified its role as a canonical testbed for methods in nonlinear science, control theory, and numerical analysis.

1. Mathematical Formulation and Canonical Properties

The classical Mackey-Glass equation arises in models of hematopoiesis and reads, after nondimensionalization,

$$\dot{x}(t) = \alpha\, \frac{x(t-\Gamma)}{1 + x^n(t-\Gamma)} - x(t),$$

where $x(t)$ is the system state, $\Gamma$ is the delay and $n$ the nonlinearity exponent. The production term is modeled by the nonlinear function $f(P) = \beta_0 P_\tau / (\Theta^n + P_\tau^n)$ (Amil et al., 2014). This equation is infinite-dimensional, as the phase space at time $t$ must specify the system's history over $[t-\Gamma, t]$.

The dynamics depend strongly on parameters:

• For small $\Gamma$ or $n<4$ the system stabilizes at fixed points.
• For larger delay or more pronounced nonlinearity, the model exhibits bifurcations to periodic orbits and chaos (Walther, 2020).
• Multistability arises: for identical parameters but different histories, distinct attractors may coexist (Tarigo et al., 2021).

Recent generalizations incorporate fractional-order derivatives to better capture memory effects (Gupta et al., 5 Nov 2024): $$D^\alpha x(t) = -p x(t) + q\,\frac{x(t-\tau)}{1 + x^r(t-\tau)},$$ where $D^\alpha$ denotes a Caputo fractional derivative.

2. Electronic and Physical Implementations

Precise analog emulation of the Mackey-Glass equation allows experimentalists to paper complex time-delay dynamics. The benchmark circuit described in (Amil et al., 2014) decomposes the model into two hardware blocks:

• Nonlinear Function Block: Implemented using precision analog multipliers/dividers (such as AD633JN, AD712JN), with exponent $n=4$. The scaling parameters are set by resistor and trimmer networks, realizing $f(v_t) = \beta v_t / (\theta^n + v_t^n)$.
• Delay Block: Utilizes a Bucket Brigade Device (BBD) with $N$ capacitors, discretely sampling the input and holding each sample for a period $T_s$ so that the effective delay is $\tau = N T_s$. The transfer is $$v_{out}(t) = v_{in}(T_s\,\lfloor t/(T_s - N + 1)\rfloor)$$, with offset/gain corrections applied.

Experiments confirm high-fidelity reproduction of the theoretical bifurcation diagrams and waveforms for all regimes: stationary, periodic, and chaotic (Amil et al., 2014, Amil et al., 2014). The flexibility to seed arbitrary initial conditions allows the exploration of phase-space structure and multistability.

A distinct physical realization uses magnetic domain walls manipulated via spin-transfer torque and resistive feedback (Williame et al., 2021), offering a nanoscience route to time-delay oscillators with Mackey-Glass nonlinearity.

3. Bifurcation Structure, Multistability, and Dynamics

Numerical bifurcation analysis reveals the intricate landscape of Mackey-Glass dynamics:

• Hopf Bifurcation: For increasing delay, the fixed point loses stability and periodic orbits emerge (Duruisseaux et al., 2022).
• Cusp Bifurcation of Periodic Orbits: Leads to branches of folds partitioning the parameter space into regions of bistability between periodic orbits. Subsequent cascades yield bistability between periodic and chaotic attractors.
• Global Crises: Chaotic attractors are destroyed either via boundary crises (collision with basin boundary) or interior crises (sudden topological change of attractor), often associated with the emergence or loss of unstable periodic orbits.

Quantification of attractor dimensions by Lyapunov spectra reveals “valleys” of low-dimensional dynamics and “islands” of high-dimensional chaos. The dimension is given by

$$d = k + \frac{1}{|\lambda_{k+1}|} \sum_{j=1}^k \lambda_j,$$

with $\lambda_j$ denoting ordered Lyapunov exponents.

A symbolic algorithm (Amil et al., 2014) enables classification of coexisting periodic orbits based on repeating peak patterns in the time series, modulo cyclic permutations—a necessary step since many attractors share identical period and amplitude but differ in ordering.

4. Stochastic Effects and Resonance Phenomena

The stochastic Mackey-Glass equation supplements the deterministic model with additive noise: $$dx(t) = {a\, x(t-\tau)}/{1 + x^c(t-\tau)} - b x(t) + \sigma\, dW_t,$$ where $dW_t$ is a Wiener noise. In the weak nonlinearity regime, noise induces transitions among point, periodic, and chaotic attractors, generating stochastic resonance (SR) at optimal $\sigma^*$ (Kojima et al., 8 May 2025). The numerically observed resonant period deviates from the classical $T \approx \tau$, corrected by a theoretically derived $\epsilon$: $$T = \tau (1 + \epsilon),\quad \epsilon = 1/(b\tau + \ln(a\tau) - 1),$$ suggesting a dominant frequency set by linear modes of an unstable spiral.

In strong nonlinearity, novel high-dimensional chaotic SR arises with multiple positive Lyapunov exponents and traveling wave patterns in the delay-memory space. The resonance point may occur before the largest Lyapunov exponent crosses zero, highlighting the coexistence of structured resonance and stochastic chaos.

5. Control Strategies and Practical Applications

Chaos control in the Mackey-Glass system can be achieved via several mechanisms (Kiss et al., 2017, Gupta et al., 5 Nov 2024):

• Constant Perturbation and Proportional Feedback: Appropriately chosen additive or proportional terms force solutions into a domain of monotonic feedback, guaranteeing global convergence to equilibrium or periodic orbit and precluding chaos by classic Poincaré-Bendixson-type theorems.
• Pyragas Control (Delayed Feedback): Adding $k[x(t-\tau) - x(t)]$ regularizes the nonlinearity, making $f(\xi) + k \xi$ monotonic for suitable $k$.
• State-dependent Delays: Decreasing the delay when the system approaches critical thresholds suppresses chaotic response.
• Fractional-order Feedback: Linear feedback for the fractional Mackey-Glass model $\left(D^\alpha x(t) = \dots + k x(t)\right)$ can stabilize the desired equilibrium over parameter ranges specified by analytical criteria (Gupta et al., 5 Nov 2024).

Applications include:

• Secure communication and chaos-based signal processing exploiting the robust generation of complex temporal patterns.
• Experimental platforms for validating theoretical and numerical methods in infinite-dimensional nonlinear systems.
• Multistate storage and digital coding leveraging attractor coexistence (Tarigo et al., 2021).
• Analog, neuromorphic, and spintronic information processing architectures based on time-delay nonlinear elements (Williame et al., 2021).

6. Networks, Relay Nonlinearities, and Further Generalizations

Extending the benchmark to fully coupled networks yields relay-generator systems, where in the limit $n \to \infty$ the nonlinearity acts as a relay with

$$F(w) = \begin{cases} 1, &\text{if}\ 0 < w < 1; \ 0, &\text{otherwise.} \end{cases}$$

Discrete traveling wave solutions $u_j(t) = u(t + \theta_j)$ emerge, with the periodic function $u(t)$ constructed to possess minimal switching points in its analytical form (Alekseev et al., 21 Jun 2024). The number of such waves grows factorially with the number $m+1$ of network generators, indicating extreme multistability.

Fractional-order, two-term, and other functional generalizations further enrich the bifurcation structure, stability phenomena, and routes to chaos, and enable more accurate modeling of nonlocal effects in natural systems (Gupta et al., 5 Nov 2024).

7. Historical Role and Mathematical Contributions

Since its publication in 1977, the Mackey-Glass equation has catalyzed rigorous advances in delay differential equations, including:

• Bifurcation theory for infinite-dimensional semiflows, Hopf and cusp bifurcations, and multiple periodic orbits (Walther, 2020, Duruisseaux et al., 2022).
• Reduction methods projecting infinite-dimensional dynamics onto finite-dimensional maps for analysis of Li-Yorke chaos.
• Floquet theory and boundary value characterizations of secondary bifurcations.
• Rich connections with physiological models, chaos theory, time-series forecasting, and high-dimensional nonlinear prediction (Walther, 2020, Walther, 2020).

The benchmark remains indispensable for validating analytic, algorithmic, experimental, and numerical methods in time-delay nonlinear systems.

---

In summary, the Mackey–Glass benchmark is defined by its infinite-dimensional delayed feedback structure, canonical nonlinear response, and multistability, and remains foundational in theoretical, experimental, and applied nonlinear science. Its diverse implementations, bifurcation phenomena, resonance effects, chaos control strategies, and generalizations enable researchers to probe and operationalize complex dynamical behavior across disciplines.