Fact-checked by Grok 2 weeks ago

Mittag-Leffler function

The Mittag-Leffler function is a family of special functions in mathematics, primarily used in complex analysis and fractional calculus, defined for a complex variable z and parameter \alpha > 0 by the power series E_\alpha(z) = \sum_{k=0}^\infty \frac{z^k}{\Gamma(\alpha k + 1)}, where \Gamma denotes the gamma function.^[1] This function generalizes the exponential function, as E_1(z) = e^z, and exhibits asymptotic behaviors that interpolate between exponential growth and power-law decay, making it essential for modeling non-local and memory-dependent phenomena.^[1] A two-parameter generalization, E_{\alpha,\beta}(z) = \sum_{k=0}^\infty \frac{z^k}{\Gamma(\alpha k + \beta)} for \beta > 0, extends its utility in solving more diverse equations.^[1] The function was introduced by the Swedish mathematician Gösta Mittag-Leffler in 1903 as part of his work on the analytic representation of functions near singular points and the summation of divergent series.^[1] Mittag-Leffler, born in 1846 and a prominent figure in late 19th- and early 20th-century mathematics, published several papers on the topic between 1902 and 1905 in outlets such as Comptes Rendus de l'Académie des Sciences and Rendiconti del Circolo Matematico di Palermo.^[1] The two-parameter form was later developed by Adolf Wiman in 1905 to study the growth of entire functions.^[1] Initially motivated by problems in function theory, the function's significance grew in the mid-20th century through connections to fractional calculus, as highlighted in works like Erdélyi et al.'s Higher Transcendental Functions (1955).^[2] Key properties of the Mittag-Leffler function include its status as an entire function of order $1/\alpha, with zeros distributed in specific sectors of the complex plane, and various integral representations, such as E_\alpha(z) = \frac{1}{2\pi i} \oint \frac{t^{\alpha-1} e^t}{t^\alpha - z} dt along a suitable contour.^[1] It satisfies differential equations like the fractional relaxation equation D^\alpha y(t) = -y(t) with solution y(t) = E_\alpha(-t^\alpha), where D^\alpha is the Caputo derivative.^[2] For $0 < \alpha < 1, E_\alpha(-t^\alpha) is completely monotone, linking it to probability theory and Laplace transforms.^[1] Applications of the Mittag-Leffler function span fractional differential equations in physics and engineering, where it provides explicit solutions to models of viscoelasticity, anomalous diffusion, and electromagnetic wave propagation in complex media.^[2] In stochastic processes, it appears in the probability densities of subordinated Lévy processes and renewal theory for non-Markovian systems.^[1] More recent extensions, such as multi-parameter variants, have been applied in quantum mechanics and image processing for handling fractional-order operators.^[2]

History and Definition

Historical Background

The Mittag-Leffler function was introduced by Gösta Mittag-Leffler, a leading Swedish mathematician renowned for founding the journal Acta Mathematica in 1882, which became a cornerstone of international mathematical publishing. In 1903, he presented the one-parameter version of the function in a seminal note to the French Academy of Sciences, marking its debut in the mathematical literature as a novel entire function generalizing classical transcendental functions.^[3] Mittag-Leffler's primary motivation for developing the function stemmed from problems in complex function theory, including the analytic representation of functions near singular points and the summation of divergent series. This generalization provided a more flexible framework within complex analysis that extended classical approaches.^[3]^[4] In 1905, the two-parameter generalization was introduced by Adolf Wiman in a paper published in Acta Mathematica, further refining its analytic properties and establishing its place within complex analysis. This work received early recognition among analysts for its contributions to the representation of multivalued functions and their uniform branches. Mittag-Leffler also contributed to the study of the function in subsequent papers during this period.^[3] Despite these foundational advances, the function remained relatively obscure in mainstream mathematics for much of the 20th century, with limited exploration beyond specialized analytic contexts. Its significance revived in the 1990s alongside the resurgence of fractional calculus applications in physics, engineering, and other sciences, prompting renewed interest in its properties and extensions.^[4]^[3]

Definition and Parameters

The Mittag-Leffler function, denoted as E_\alpha(z), is a special function defined for complex variables z \in \mathbb{C} and parameter \alpha \in \mathbb{C} by the power series

E_\alpha(z) = \sum_{k=0}^\infty \frac{z^k}{\Gamma(\alpha k + 1)},

where \Gamma denotes the gamma function. This one-parameter form was originally introduced by Gösta Mittag-Leffler in 1903 as a generalization of the exponential function in the context of entire functions. The parameter \alpha governs the growth rate and order of the function, with the series serving as its defining representation.^[5] A two-parameter generalization, E_{\alpha,\beta}(z), extends this definition to include an additional complex parameter \beta \in \mathbb{C}:

E_{\alpha,\beta}(z) = \sum_{k=0}^\infty \frac{z^k}{\Gamma(\alpha k + \beta)}.

Here, \beta shifts the argument in the gamma function denominator, influencing the initial terms and overall behavior of the series while preserving the role of \alpha in controlling growth. This form, introduced by Adolf Wiman in 1905, is entire (analytic everywhere in the complex plane) when \operatorname{Re}(\alpha) > 0 and \operatorname{Re}(\beta) > 0. Notation may vary, with subscripts \alpha, \beta or a, b appearing in literature, but the standard E_{\alpha,\beta}(z) emphasizes the parametric dependence.^[5] Regarding convergence, the series for both forms exhibits absolute convergence for all z \in \mathbb{C} when \operatorname{Re}(\alpha) > 0, yielding an infinite radius of convergence and rendering the functions entire. For \operatorname{Re}(\alpha) \leq 0, convergence is restricted to a disk of finite radius in the complex plane, determined by the asymptotic growth of the gamma function terms \Gamma(\alpha k + \beta), which no longer decay sufficiently fast for large |z|. These conditions ensure the function's utility in applications requiring global analyticity.^[6]

Properties

Analytic Continuation and Order

The Mittag-Leffler function E_{\alpha,\beta}(z), defined for parameters with \operatorname{Re}(\alpha) > 0 and \operatorname{Re}(\beta) > 0, is an entire function of the complex variable z, meaning it is holomorphic everywhere in the finite complex plane. This property follows directly from the power series representation

E_{\alpha,\beta}(z) = \sum_{k=0}^{\infty} \frac{z^k}{\Gamma(\alpha k + \beta)},

which converges absolutely and uniformly on every compact subset of \mathbb{C}, establishing its analyticity throughout the plane.^[7] As an entire function, E_{\alpha,\beta}(z) has order \rho = 1/\alpha and type \sigma = 1. The order \rho characterizes the overall growth rate, with \rho < \infty ensuring finite order, while the type \sigma quantifies the precise exponential growth along directions of maximal increase, specifically \max_{|z|=r} \log |E_{\alpha,\beta}(z)| \sim \sigma r^{\rho} as r \to \infty. This type achieves its minimum value of 1 when \alpha = 1, where E_{1,\beta}(z) reduces to a form related to the exponential function scaled by the gamma function. For $0 < \alpha < 1, the order exceeds 1, reflecting slower decay in certain directions compared to the exponential case.^[8] The analytic continuation of E_{\alpha,\beta}(z) is inherently provided by the series for the specified parameter range, requiring no additional constructions beyond the disk of convergence, as the radius is infinite. However, for \operatorname{Re}(\alpha) \leq 0, the function generally requires branch cuts to define a multi-valued continuation, often along the negative real axis, due to singularities in the gamma function terms. Growth estimates within the unit disk, where |z| < 1, satisfy |E_{\alpha,\beta}(z)| \leq M / (1 - |z|) for some constant M > 0 depending on \alpha and \beta, derived from the dominated convergence of the series terms.^[7] Applications of the Phragmén–Lindelöf principle to E_{\alpha,\beta}(z) highlight its maximal angular domains of bounded growth, particularly in sectors where |\arg z| \leq \pi \alpha / 2, ensuring that the function remains controlled in these regions despite exponential growth elsewhere. This principle aids in bounding the function in unbounded sectors, confirming the sharpness of the order and type estimates. The entire nature of the function also implies uniqueness in the sense of meromorphic continuation theorems, though its pole-free structure distinguishes it from typical Mittag-Leffler theorem applications to functions with prescribed singularities.^[9]^[10]

Recurrence and Differentiation

The Mittag-Leffler function satisfies a fundamental differentiation formula that relates its derivative to other Mittag-Leffler functions with shifted parameters. Specifically, the first-order derivative is given by

\frac{d}{dz} E_{\alpha,\beta}(z) = \frac{E_{\alpha,\beta-1}(z) + (1 - \beta) E_{\alpha,\beta}(z)}{\alpha z},

valid for z \neq 0 and parameters \alpha > 0, \beta \in \mathbb{C} with \operatorname{Re}(\beta) > 0.^[11] This relation can be derived from the series expansion of E_{\alpha,\beta}(z) by term-by-term differentiation.^[11] A closely related recurrence relation connects the derivative to a linear combination of Mittag-Leffler functions without the $1/z factor:

\alpha z \frac{d}{dz} E_{\alpha,\beta}(z) = E_{\alpha,\beta-1}(z) - (\beta - 1) E_{\alpha,\beta}(z).

This form highlights the operational similarity to the exponential function, where \alpha = 1, \beta = 1 reduces it to the standard product rule.^[12] The relation extends iteratively to higher-order derivatives. For the m-th derivative, one expression involves a finite sum:

(\alpha z)^m \frac{d^m}{dz^m} E_{\alpha,\beta}(z) = \sum_{j=0}^m q_j(\alpha, \beta, m) E_{\alpha, \beta - j}(z),

where the coefficients q_j(\alpha, \beta, m) are polynomials in the parameters, with explicit forms for low j such as q_m = 1 and q_0 = (-1)^m \prod_{k=0}^{m-1} (\beta - 1 + k \alpha).^[11]^[12] These recurrences facilitate numerical computation and analysis in applications involving repeated differentiation.^[11] In the context of fractional calculus, the Mittag-Leffler function exhibits invariance-like properties under fractional differentiation operators. For the Caputo fractional derivative of order \gamma (with $0 < \gamma < 1), applied to the form t^{\beta-1} E_{\alpha,\beta}(\lambda t^\alpha), the result is

{}^C D^\gamma \left[ t^{\beta-1} E_{\alpha,\beta}(\lambda t^\alpha) \right] = t^{\beta - \gamma - 1} E_{\alpha, \beta - \gamma}(\lambda t^\alpha),

provided \operatorname{Re}(\beta) > \gamma and other standard conditions on the parameters hold to ensure convergence.^[13] This formula underscores the function's role as a natural kernel in fractional differential equations, where fractional derivatives preserve the structural form up to parameter shifts.^[13] Similar expressions exist for the Riemann-Liouville derivative, though they include additional terms depending on initial conditions.^[13] An analogous relation for ordinary differentiation provides a basis for integration by parts specific to the Mittag-Leffler function:

\frac{d}{dz} \left[ z^{\beta-1} E_{\alpha,\beta}(z^\alpha) \right] = z^{\beta-2} E_{\alpha,\beta-1}(z^\alpha).

Integrating both sides yields the antiderivative in terms of a shifted Mittag-Leffler function, enabling efficient handling of integrals in series solutions or transform methods without expanding the full series.^[12] This property is particularly useful for deriving moment formulas or resolving integral equations involving the function.^[12]

Representations

Series Expansion

The Mittag-Leffler function E_{\alpha,\beta}(z) is fundamentally defined by its power series expansion, known as the Maclaurin series, which serves as the primary tool for its analytic representation. For complex parameters \alpha with \operatorname{Re}(\alpha) > 0 and \beta \in \mathbb{C}, and z \in \mathbb{C}, the series is given by

E_{\alpha,\beta}(z) = \sum_{k=0}^{\infty} \frac{z^k}{\Gamma(\beta + \alpha k)},

where \Gamma denotes the gamma function. This expansion arises directly from the function's construction as a generalization of the exponential series, where the factorial denominators are replaced by gamma function terms to accommodate fractional orders. The first few terms illustrate its structure: E_{\alpha,\beta}(0) = 1/\Gamma(\beta), and higher terms incorporate the scaling by \alpha in the argument of the gamma function. The series converges absolutely for all finite z \in \mathbb{C}, making E_{\alpha,\beta}(z) an entire function when \operatorname{Re}(\alpha) > 0. To establish this, apply the ratio test to the coefficients a_k = 1 / \Gamma(\beta + \alpha k): the ratio |a_{k+1}/a_k| = \Gamma(\beta + \alpha k) / \Gamma(\beta + \alpha (k+1)). Using the asymptotic behavior of the gamma function for large k, \Gamma(w + \alpha) / \Gamma(w) \sim w^{\alpha}, so |a_{k+1}/a_k| \sim 1 / [(\alpha k)^{\alpha}]. Thus, \lim_{k \to \infty} |a_{k+1} z / a_k| = 0 for any fixed z, confirming an infinite radius of convergence. The order of the entire function is \rho = 1/\alpha. For practical computation, the series is truncated to a finite partial sum E_{\alpha,\beta}(z) \approx \sum_{k=0}^{N} z^k / \Gamma(\beta + \alpha k) when |z| is small, as the terms decay rapidly due to the growth of the gamma function in the denominator. This approach is efficient for |z| \lesssim 1, with the truncation error bounded by the remainder term, which can be estimated using integral representations of the gamma function or direct summation until terms fall below machine precision. For larger |z|, alternative methods like asymptotic expansions are preferred to avoid slow convergence.^[14] The Mittag-Leffler function is a special case of the generalized Wright function with p=1, specifically E_{\alpha,\beta}(z) = {}_1 \Psi_1 \left[ \begin{array}{c} (1,1) \\ (\beta, \alpha) \end{array} ; z \right], where the Wright function generalizes hypergeometric series through products of gamma functions in the numerator and denominator. The (1,1) parameter in the numerator cancels the k! denominator present in the general Wright series definition.

Integral and Contour Representations

The Mittag-Leffler function admits several integral and contour representations that facilitate its analytic continuation, asymptotic evaluation, and application in solving integral equations. These representations often leverage contour integrals in the complex plane to express the function in closed form, particularly useful when the series expansion converges slowly or for large |z|. A key example is the Hankel contour integral, obtained by substituting the Hankel representation of the reciprocal gamma function into the series definition of E_{\alpha,\beta}(z).^[15] The Hankel contour integral representation is given by

E_{\alpha,\beta}(z) = \frac{1}{2\pi i} \int_H \frac{e^t (-t)^{-\beta}}{1 - z (-t)^{-\alpha}} \, dt,

where the Hankel contour H originates at -\infty just below the negative real axis, encircles the origin counterclockwise in a small circle of radius \epsilon > 0, and returns to -\infty just above the negative real axis, with the principal branch of the multivalued functions defined by -\pi < \arg(-t) < \pi. This representation is valid for \operatorname{Re}(\beta) > 0, $0 < \alpha < 2, and | \arg z | < \min(\pi, \alpha \pi / 2), ensuring the geometric series inside the integral converges and the contour avoids branch cuts.^[15] Another fundamental representation arises from the Laplace transform, which connects the Mittag-Leffler function to solutions of fractional differential equations. Specifically, the Laplace transform of t^{\beta-1} E_{\alpha,\beta}(t^\alpha) is

\mathcal{L}\{ t^{\beta-1} E_{\alpha,\beta}(t^\alpha) \}(s) = \frac{s^{\alpha - \beta}}{s^\alpha - 1},

valid for \operatorname{Re}(s) > 0 and under conditions ensuring the transform exists, such as $0 < \alpha \leq 1 and \beta > 0. More generally, for E_{\alpha,\beta}(\lambda t^\alpha), the formula becomes \mathcal{L}\{ t^{\beta-1} E_{\alpha,\beta}(\lambda t^\alpha) \}(s) = s^{\alpha - \beta} / (s^\alpha - \lambda), with \operatorname{Re}(s) > |\lambda|^{1/\alpha}. This relation, derived via term-by-term integration of the series, underscores the function's role as a kernel in fractional calculus.^[15] The Mittag-Leffler function also appears as a special case of the Fox–Wright function, a generalized hypergeometric function defined by the series

\,^p\Psi_q \left( z \,\middle|\, \begin{array}{c} (a_i, A_i)_{i=1}^p \\ (b_j, B_j)_{j=1}^q \end{array} \right) = \sum_{k=0}^\infty \frac{\prod_{i=1}^p \Gamma(a_i + A_i k)}{\prod_{j=1}^q \Gamma(b_j + B_j k)} \frac{z^k}{k!}

. Specifically,

E_{\alpha,\beta}(z) = \,^1\Psi_1 \left( z \,\middle|\, \begin{array}{c} (1, 1) \\ (\beta, \alpha) \end{array} \right)

. The Fox–Wright function possesses a Mellin–Barnes contour integral representation:

\,^p\Psi_q \left( z \,\middle|\, \begin{array}{c} (a_i, A_i)_{i=1}^p \\ (b_j, B_j)_{j=1}^q \end{array} \right) = \frac{1}{2\pi i} \int_L \frac{\prod_{i=1}^p \Gamma(a_i + A_i s) \prod_{j=1}^q \Gamma(1 - b_j - B_j s)}{\prod_{i=1}^p \Gamma(1 - a_i - A_i s) \prod_{j=1}^q \Gamma(b_j + B_j s)} (-z)^{-s} \, ds,

where L is a vertical contour in the complex s-plane separating the poles of the gamma functions in the numerator from those in the denominator, chosen such that the integral converges for appropriate z. This form directly applies to the Mittag-Leffler function via the parameter specialization, providing a powerful tool for analytic continuation.^[15] For multi-parameter extensions, such as the Prabhakar function (a three-parameter generalization), the Mellin–Barnes representation extends naturally by incorporating additional gamma factors in the integrand, maintaining the contour structure while adjusting the pole locations for the extra parameters. This allows evaluation of generalized Mittag–Leffler functions in broader domains.^[15]

Special Cases

Reduction to Exponential and Error Functions

The Mittag-Leffler function E_{\alpha,\beta}(z) admits exact reductions to elementary functions or the error function for certain rational values of the parameters \alpha and \beta, particularly when \alpha = 1, \alpha = 1/2, or \alpha = 2. These cases arise naturally from the power series definition E_{\alpha,\beta}(z) = \sum_{k=0}^{\infty} \frac{z^k}{\Gamma(\alpha k + \beta)}, where the denominator aligns with the series expansions of exponential, hyperbolic, or error-related functions. When \alpha = 1 and \beta = 1, the function simplifies directly to the exponential function:

E_{1,1}(z) = e^z.

This follows from the series \sum_{k=0}^{\infty} \frac{z^k}{k!} = e^z, establishing the Mittag-Leffler function as a generalization of the exponential. For \alpha = 1 and general \beta > 0, E_{1,\beta}(z) = \frac{{}_1F_1(1;\beta;z)}{\Gamma(\beta)}, where {}_1F_1 is Kummer's confluent hypergeometric function.^[6] For the case \beta = 1 and \alpha = 1/2, the one-parameter Mittag-Leffler function E_{1/2}(z) = E_{1/2,1}(z) reduces to a form involving the complementary error function \operatorname{erfc}(w) = 1 - \operatorname{erf}(w) = \frac{2}{\sqrt{\pi}} \int_w^{\infty} e^{-t^2} \, dt:

E_{1/2,1}(\sqrt{z}) = e^z \, \operatorname{erfc}(-\sqrt{z}).

Equivalently, in terms of the argument z,

E_{1/2,1}(z) = e^{z^2} \, \operatorname{erfc}(-z).

This reduction highlights the connection to diffusion processes, where the error function appears in Gaussian integrals, and the quadratic exponent reflects the half-order fractional dynamics.^[16] When \beta = 1 and \alpha = 2, the function ties to hyperbolic functions via the series alignment with even powers:

E_{2,1}(z) = \cosh(\sqrt{z}).

This identity stems from \sum_{k=0}^{\infty} \frac{z^k}{\Gamma(2k + 1)} = \sum_{k=0}^{\infty} \frac{(\sqrt{z})^{2k}}{(2k)!} = \cosh(\sqrt{z}). Variants for other \beta with \alpha = 2 include

E_{2,2}(z) = \frac{\sinh(\sqrt{z})}{\sqrt{z}},

derived similarly from the series for \sinh(u)/u = \sum_{k=0}^{\infty} \frac{u^{2k}}{(2k+1)!}, with u = \sqrt{z} and \Gamma(2k + 2) = (2k + 1)!. These hyperbolic reductions are particularly useful in wave propagation models, where \alpha = 2 corresponds to second-order behavior akin to classical oscillators.

Other Limiting and Fractional Cases

As the parameter α approaches 0 from above, the Mittag-Leffler function E_{α,β}(z) converges to E_{0,β}(z) = \frac{1}{\Gamma(β)(1-z)} for |z| < 1. This limiting case arises from the power series definition, where each term z^k / Γ(αk + β) approaches z^k / Γ(β), yielding a geometric series scaled by 1/Γ(β). When α is an integer greater than 1, the Mittag-Leffler function admits expressions in terms of elementary or other special functions. For instance, with α = n ≥ 2 and β = 1,

E_{n,1}(z) = \frac{1}{n} \sum_{k=0}^{n-1} \exp\left( z^{1/n} \exp\left( \frac{2 \pi i k}{n} \right) \right).

For general integer α = n > 1 and β, relations often involve confluent hypergeometric or other special functions. These connections are useful for asymptotic analysis and applications in wave propagation models. In the asymptotic regime for large α with fixed z, the Mittag-Leffler function E_α(z) approaches 1. For the large |z| asymptotic in the sector |arg z| < (α π)/2 with 0 < α < 2, E_α(z) \sim \frac{1}{\alpha} z^{(1 - \alpha)/\alpha} \exp(z^{1/\alpha}). For α ≥ 2, additional terms from other sectors contribute via roots of unity in the exact integer case, reflecting a transition toward integer-order dynamics dominated by exponentials. This large-α limit underscores the function's generalization of the exponential, with integral representations aiding computational verification for practical use.^[16]

Generalizations

Two-Parameter Extension

The two-parameter Mittag-Leffler function extends the original one-parameter form E_\alpha(z) = E_{\alpha,1}(z) by incorporating an additional parameter \beta \in \mathbb{C} that shifts the argument in the gamma function within the series expansion, thereby broadening its applicability in fractional differential equations and related fields. Introduced by Wiman in 1905 as a means to study the zeros of certain entire functions, this generalization is defined for \alpha > 0 and z \in \mathbb{C} by the power series

E_{\alpha,\beta}(z) = \sum_{k=0}^\infty \frac{z^k}{\Gamma(\alpha k + \beta)},

which converges absolutely in the entire complex plane under appropriate conditions on the parameters. This modification of the denominator from \Gamma(\alpha k + 1) to \Gamma(\alpha k + \beta) enables the function to capture more nuanced decay and oscillatory behaviors, particularly when \beta \neq 1. A key property distinguishing the two-parameter version is its compatibility with fractional integral operators, such as the Riemann-Liouville fractional integral of order \gamma > 0, defined as

{}^RL I_0^\gamma f(t) = \frac{1}{\Gamma(\gamma)} \int_0^t (t - s)^{\gamma - 1} f(s) \, ds.

For f(t) = t^{\beta - 1} E_{\alpha,\beta}(\lambda t^\alpha) with \operatorname{Re}(\beta) > 0 and \lambda \in \mathbb{C}, it holds that

{}^RL I_0^\gamma \left[ t^{\beta - 1} E_{\alpha,\beta}(\lambda t^\alpha) \right] = t^{\beta + \gamma - 1} E_{\alpha,\beta + \gamma}(\lambda t^\alpha).

When \gamma = 1, this specializes to the ordinary integral relation

\int_0^t s^{\beta - 1} E_{\alpha,\beta}(\lambda s^\alpha) \, ds = t^\beta E_{\alpha,\beta + 1}(\lambda t^\alpha).

These relations underscore the function's role in solving fractional integral equations, as they preserve the Mittag-Leffler structure under integration. Regarding convergence and analyticity, the power series defines an entire function of order \rho = 1/\alpha < \infty and type depending on |z| when \operatorname{Re}(\alpha) > 0 and \operatorname{Re}(\beta) > 0, ensuring no poles in the terms since \alpha k + \beta avoids the non-positive integers where \Gamma has poles. For \operatorname{Re}(\beta) \leq 0, the series may diverge at specific points due to poles in the gamma function; for instance, when \beta = 0, the k=0 term involves $1/\Gamma(0), which is undefined, necessitating analytic continuation via limits or integral representations to extend the function beyond its principal branch. Similarly, for negative integer values like \beta = -m with m \in \mathbb{N}, additional poles arise for small k where \alpha k - m hits non-positive integers, but the function remains meromorphic in such cases after continuation, with residues computable from the gamma function properties. The two-parameter Mittag-Leffler function also admits a representation as a special case of the Wright function, a generalized hypergeometric function of one upper and one lower parameter set, given by

E_{\alpha,\beta}(z) = \frac{1}{\Gamma(\beta)} {}_1\Psi_1 \left[ \begin{matrix} (1,1) \\ (\beta, \alpha) \end{matrix} ; z \right] = \frac{1}{\Gamma(\beta)} \sum_{k=0}^\infty \frac{(1)_k}{(\beta)_{\alpha k}} \frac{z^k}{k!},

where ( \cdot )_ \cdot denotes the Pochhammer symbol; this connection highlights its place within the hierarchy of confluent hypergeometric functions and facilitates asymptotic expansions for large |z|.

Prabhakar and Multi-Parameter Variants

The Prabhakar function, also known as the three-parameter Mittag-Leffler function, generalizes the two-parameter form by incorporating a third parameter that introduces a Pochhammer symbol in the series expansion. Defined as

E_{\alpha,\beta}^{\gamma}(z) = \sum_{k=0}^{\infty} \frac{(\gamma)_k z^k}{k! \Gamma(\alpha k + \beta)},

where (\gamma)_k denotes the rising Pochhammer symbol and \alpha, \beta, \gamma > 0 with z \in \mathbb{C}, it was introduced by Tilak Raj Prabhakar in 1971 to solve a singular integral equation.^[17] This function reduces to the standard two-parameter Mittag-Leffler function when \gamma = 1. Key properties of the Prabhakar function include its role as a kernel in fractional calculus operators, where the Prabhakar integral is given by I^{\alpha,\beta,\gamma} f(t) = \int_0^t (t - \tau)^{\beta - 1} E_{\alpha,\beta}^{\gamma}(\lambda (t - \tau)^\alpha) f(\tau) \, d\tau, enabling the formulation of non-local fractional derivatives with memory effects beyond the classical Riemann-Liouville or Caputo types.^[18] Its Laplace transform is particularly useful for analyzing solutions to fractional differential equations: \mathcal{L}\{ t^{\beta-1} E_{\alpha,\beta}^{\gamma}(\omega t^\alpha) \}(s) = s^{\alpha\gamma - \beta} / (s^\alpha - \omega)^\gamma for \operatorname{Re}(s) > 0 and appropriate \omega.^[19] Additionally, the function contributes to stability analysis in Volterra-type integral equations of fractional order, where solutions exhibit Mittag-Leffler-type decay bounds ensuring asymptotic stability under suitable conditions on the parameters.^[20] Recent extensions of the Prabhakar function have emerged to address more complex models in fractional dynamics, as of 2025. The g-generalized Mittag-Leffler (p,s,k)-function incorporates a g-parameter to modulate the series coefficients, enhancing flexibility for multi-scale phenomena while preserving analytic continuation properties.^[21] In 2025, q-analogues of the Mittag-Leffler function, such as the generalized q-Mittag-Leffler function, were developed using q-calculus to discretize the continuous framework, with applications in quantum statistics; for instance, the generalized form replaces factorials with q-factorials and uses q-Pochhammer symbols.^[22] The extended k-generalized form further broadens this by integrating extensions via generalized beta functions, allowing for tunable asymptotic behaviors in the complex plane.^[23] Multi-parameter variants, such as those discussed by Kilbas et al., extend the forms to include additional gamma functions in the denominator, e.g., E_{\alpha_1,\beta_1;\alpha_2,\beta_2}(z) = \sum_{k=0}^{\infty} \frac{z^k}{\Gamma(\alpha_1 k + \beta_1) \Gamma(\alpha_2 k + \beta_2)}, facilitating representations in quantum mechanical systems through their connections to hypergeometric series and integral transforms.^[24]

Applications

In Fractional Calculus

The Mittag-Leffler function plays a central role in solving fractional differential equations of relaxation type, where the Caputo fractional derivative of order \alpha \in (0,1) is involved. Consider the fractional relaxation equation {}^C D^\alpha y(t) = -\lambda y(t) with \lambda > 0, which generalizes the classical exponential decay. The solution, with initial condition y(0) = y_0, takes the form y(t) = y_0 E_\alpha (-\lambda t^\alpha), where E_\alpha(z) is the one-parameter Mittag-Leffler function.^[25] This expression captures the power-law decay and intermediate asymptotic behavior characteristic of fractional systems, distinguishing it from the exponential solution in integer-order cases.^[26] In the context of stability analysis for linear fractional systems using the Caputo derivative, the concept of Mittag-Leffler stability emerges as a refinement of asymptotic stability. For the equation {}^C D^\alpha y(t) = \lambda y(t), the origin is Mittag-Leffler stable if the solution satisfies |y(t)| \leq C t^{-\alpha} for some constant C > 0, reflecting the slower decay compared to exponential rates. Asymptotic decay holds when |\arg(\lambda)| > \alpha \pi / 2, ensuring the eigenvalues lie outside the principal stability sector in the complex plane; this condition guarantees that the Mittag-Leffler function in the solution decays to zero as t \to \infty.^[27]^[28] The Mittag-Leffler function also appears prominently in the theory of Volterra integral equations of fractional type. For the Volterra equation of the second kind y(t) = f(t) + \lambda \int_0^t (t-s)^{\alpha-1} y(s) \, ds / \Gamma(\alpha), the resolvent kernel is given by (t-s)^{\alpha-1} E_{\alpha,\alpha} (\lambda (t-s)^\alpha), allowing the solution to be expressed as a convolution: y(t) = f(t) + \lambda \int_0^t (t-s)^{\alpha-1} E_{\alpha,\alpha} (\lambda (t-s)^\alpha) f(s) \, ds. This structure facilitates the analysis of existence, uniqueness, and regularity of solutions in fractional integral frameworks.^[25] Recent developments in fractional differential equations (2020s) have leveraged the Mittag-Leffler function in modeling anomalous diffusion processes, where solutions to time-fractional diffusion equations exhibit non-Gaussian spreading behaviors captured by Mittag-Leffler tails. For instance, in fractional oscillator models, such as the forced fractional harmonic oscillator {}^C D^\alpha x(t) + \omega^2 x(t) = f(t), the solution involves multivariate extensions of the Mittag-Leffler function, enabling descriptions of damped oscillations with memory effects. These applications highlight the function's utility in deriving exact solutions and analyzing long-time asymptotics in literature on fractional dynamics.^[29]

In Viscoelasticity and Diffusion Models

The Mittag-Leffler function plays a central role in modeling the viscoelastic behavior of materials through fractional derivative formulations, particularly in the fractional Zener model, which captures the relaxation and creep responses of solid-like viscoelastic systems. In this model, the relaxation modulus G(t) is expressed as G(t) = G_0 E_\alpha \left( -\left( \frac{t}{\tau} \right)^\alpha \right), where G_0 is the initial shear modulus, \tau is the characteristic relaxation time, $0 < \alpha < 1 is the fractional order reflecting material memory, and E_\alpha denotes the one-parameter Mittag-Leffler function.^[30] This form arises from solving the fractional differential equation governing stress-strain relations in the model, providing a power-law decay at long times that mimics the behavior observed in rubber-like polymers under sustained loading.^[30] Such applications are prevalent in polymer science, where the function's asymptotic properties enable accurate fitting of experimental relaxation data for materials exhibiting intermediate viscoelasticity between elastic solids and viscous fluids.^[31] Complementing relaxation, the creep compliance J(t) in the fractional Zener model incorporates the two-parameter Mittag-Leffler function as J(t) = \frac{1}{b_1} t^\alpha E_{\alpha, \alpha+1} \left( -\frac{b_0}{b_1} t^\alpha \right) + \frac{a}{b_1} E_{\alpha, 1} \left( -\frac{b_0}{b_1} t^\alpha \right), with coefficients a, b_0, and b_1 derived from model parameters like spring constants and relaxation times.^[32] Here, E_{\alpha,1+\alpha} captures the retarded elastic response, leading to sublinear creep growth that aligns with observations in polymeric materials under constant stress, enhancing predictions of long-term deformation.^[32] In diffusion processes, the Mittag-Leffler function describes anomalous subdiffusion in porous media, where particle transport deviates from classical Fickian behavior due to trapping and heterogeneous structures. The probability density function for subdiffusive spread is often given by a Fox's H-function representation that reduces to forms involving E_{\alpha,\beta} for the two-parameter Mittag-Leffler function, yielding heavy-tailed distributions with mean squared displacement scaling as t^\alpha ($0 < \alpha < 1).^[33] This is particularly relevant in modeling solute transport through biological tissues or geological formations, where the function's non-exponential decay reflects prolonged residence times in pores.^[34] Recent advancements in the 2020s have extended these applications to biomedical contexts, such as drug release from cylindrical matrices, where fractional diffusion equations solved via extended Mittag-Leffler functions predict non-Fickian release profiles with initial bursts followed by sustained delivery, improving therapeutic efficacy in controlled-release systems.^[35] Similarly, in viscoelastic wave propagation, numerical schemes approximating the Mittag-Leffler function enable efficient simulations of damped wave equations in attenuating media like soft tissues, revealing frequency-dependent dispersion critical for ultrasound imaging and seismic analysis.^[36] Advanced models occasionally incorporate Prabhakar kernels, a multi-parameter generalization of the Mittag-Leffler function, for enhanced memory effects in complex viscoelastic systems.^[37]

References

[1]
https://arxiv.org/pdf/0909.0230.pdf
[2]
[PDF] Mittag-Leffler Functions, Related Topics and Applications
The study of the Mittag-Leffler function and its various generalizations has become a very popular topic in mathematics and its applications.
[3]
Why the Mittag-Leffler Function Can Be Considered the Queen ...
In this survey we stress the importance of the higher transcendental Mittag-Leffler function in the framework of the Fractional Calculus.Missing: qui ajoute fonctions transcendantes
[4]
[0909.0230] Mittag-Leffler Functions and Their Applications - arXiv
Sep 1, 2009 · In this survey paper, nearly all types of Mittag-Leffler type functions existing in the literature are presented.Missing: definition | Show results with:definition
[5]
Mittag‐Leffler Functions and Their Applications - Wiley Online Library
May 30, 2011 · 4 Mittag-Leffler G., Mittag-Leffler, Sur la representation analytiqie d'une fonction monogene (cinquieme note), Acta Mathematica. (1905) 29 ...
[6]
[PDF] The Mittag-Leffler function - arXiv
Sep 26, 2021 · We review the function theoretical properties of the Mittag-Leffler function Ea,b (z) in a self- contained manner, but also add new results; ...
[7]
Mittag-Leffler Function -- from Wolfram MathWorld
The Mittag-Leffler function (Mittag-Leffler 1903, 1905) is an entire function defined by the series E_alpha(z)=sum_(k=0)^infty(z^k)/(Gamma(alphak+1)) for alpha
[8]
[PDF] A Littlewood-Paley approach to the Mittag-Leffler function in ... - arXiv
The Mittag-Leffler function Eα,β is entire of order 1/α, when α,β > 0 (see ... Solutions to this type of equations can also be obtained using the ...
[9]
Historical Overview of the Mittag-Leffler Functions
Summary of each segment:
[10]
Phragmén-Lindelöf Theorems - jstor
Jul 28, 2003 · PHRAGMEN-LINDELOF THEOREMS. P. C. FENTON AND JOHN ROSSI ... and let Ex(z) be the Mittag-Leffler function [2, p. 50]. EA(z) is an ...
[11]
Derivatives of Mittag-Leffler functions: theory, computation and ...
Aug 29, 2025 · ... fractional differential equations (FDEs) [3, 5, 6]. They are also instrumental in the development of efficient numerical methods for solving ...<|control11|><|separator|>
[12]
[PDF] The Mittag-Leffl er function
We review the function theoretical properties of the Mittag-Leffl er function Ea;b (z) in a self- contained manner, but also add new results. Contents. 1 ...
[13]
Fractional derivatives of the generalized Mittag-Leffler functions
Nov 9, 2018 · In this paper, we derive the compositions of the fractional derivatives with the Shukla function, a four-parameter Mittag-Leffler function.
[14]
[PDF] Computation of the generalized Mittag-Leffler function and its ...
2.3 Taylor series. For z ∈ G0, i.e. for the central disk region (with r0 = 0.9), the Mittag-Leffler function is computed from truncating its Taylor series (1).
[15]
https://doi.org/10.1007/978-3-662-43930-2
[16]
https://www.hindawi.com/journals/jam/2011/298628/
[17]
T. R. Prabhakar, “A Singular Integral Equation with a Generalized ...
T. R. Prabhakar, “A Singular Integral Equation with a Generalized Mittag-Leffler Function in the Kernel,” Yokohama Mathematical Journal, Vol. 19, 1971, pp. 7-15 ...Missing: original | Show results with:original
[18]
https://www.degruyterbrill.com/document/doi/10.1515/fca-2020-0002/html
[19]
[PDF] The Prabhakar or three parameter Mittag–Leffler function - arXiv
Oct 11, 2017 · This paper is organized as follows. We first review, in Section 2, the main properties of the Prabhakar function and we present integral and ...
[20]
Fractional Fourier Transform to Stability Analysis of ... - ResearchGate
Aug 8, 2025 · This note aims to study the Mittag-Leffler-Hyers-Ulam stability of the linear and nonlinear fractional differential equations with the Prabhakar ...
[21]
The g-generalized Mittag-Leffler (p,s,k)-function - ScienceDirect
This paper introduces the generalization of ML function, hereafter referred to as the g -generalized ML ( p , s , k ) -function, which extends the applicability ...
[22]
A New q‐Integral Operator Containing Generalized q‐Mittag‐Leffler ...
Jan 20, 2025 · In this study, we employ the mathematical framework of q-calculus to introduce a novel integral operator that incorporates the generalized q-Mittag-Leffler ...
[23]
On extended -generalized Mittag-Leffler function and its properties
The Mittag-Leffler function is being studied due to its wide area of applications like applications in solving fractional differential and integral equations, ...
[24]
Multi-parametric mittag-leffler functions and their extension
Mar 19, 2013 · This paper surveys one of the last contributions by the late Professor Anatoly Kilbas (1948–2010) and research made under his advisorship.Missing: Saxon | Show results with:Saxon
[25]
The Role of the Mittag-Leffler Function in Fractional Modeling - MDPI
This is a survey paper illuminating the distinguished role of the Mittag-Leffler function and its generalizations in fractional analysis and fractional ...Missing: obscurity revival
[26]
On Mittag-Leffler-type functions in fractional evolution processes
The purpose of this review is to outline the fundamental role of Mittag-Leffler-type functions in fractional evolution processes.
[27]
Mittag–Leffler stability of fractional order nonlinear dynamic systems
In this paper, we extend the Lyapunov direct method by considering ... Mittag-leffler stability of fractional order nonlinear dynamic systems. In ...Missing: paper | Show results with:paper
[28]
On asymptotic properties of solutions to fractional differential equations
Apr 15, 2020 · The Caputo fractional derivative of a d ... Now we show that the Mittag-Leffler stability is strictly stronger than asymptotic stability.
[29]
Multivariate Mittag-Leffler Solution for a Forced Fractional-Order ...
May 11, 2024 · The Mittag-Leffler function is a generalization of the exponential function, that is, e z = E 1 , 1 ( z ) . Extensions to two or three ...
[30]
Two Integral Representations for the Relaxation Modulus of ... - MDPI
One of the most used fractional viscoelastic models of solid-like behavior is the fractional Zener model, which is defined by the following stress–strain ...
[31]
Notes on computational aspects of the fractional-order viscoelastic ...
Jun 8, 2017 · Among these models, the fractional Zener model was found ... Mittag-Leffler function is also known as the fractional exponential function.
[32]
Simultaneous Characterization of Relaxation, Creep, Dissipation ...
Apr 20, 2021 · The relaxation modulus and the creep compliance for the six-parameter fractional constitutive model were given in terms of the Mittag–Leffler ...
[33]
Characterization of Anomalous Diffusion in Porous Biological ...
In the CTRW approach, the Fourier transform yields a solution to the generalized diffusion equation that can be expressed by the Mittag-Leffler function (MLF), ...
[34]
Generalized fractional diffusion equations for subdiffusion in ...
Oct 26, 2017 · The Mittag-Leffler probability density is heavy tailed, which is asymptotically ψ ( t ) ∼ t − 1 − α for long times. The memory kernel of a ...
[35]
Fractionalized mathematical models for drug diffusion - ScienceDirect
The general solution of drug concentration is demonstrated in the form of extended Mittag-Leffler function. The amount of drug that is contained in each ...
[36]
Analysis of a fast fully discrete finite element method for fractional ...
Sep 25, 2025 · This paper is devoted to a numerical analysis of a fractional viscoelastic wave propagation model that generalizes the fractional Maxwell model ...<|control11|><|separator|>
[37]
wave propagation in viscoelastic half-space with memory functions ...
Aug 6, 2025 · They are expressed in terms of functions of Mittag-Leffler type, including the recently introduced multinomial Prabhakar type function.