Fact-checked by Grok 2 weeks ago

Exp

In , the , commonly denoted as exp(x) or e^x, is a fundamental defined as the unique real-valued that maps zero to one and whose equals itself at every point, representing or decay proportional to the 's current value. It is specifically given by exp(x) = e^x, where e is Euler's number, the base of the natural logarithm, approximately equal to 2.718281828459. The exp(x) has a of all real numbers and a range of all , serving as the of the natural logarithm ln(x). One of the defining properties of exp(x) is its series expansion: exp(x) = ∑_{n=0}^∞ x^n / n!, which converges for all real x and provides a way to compute its values without relying on the base e directly. This highlights its smoothness and analytic nature, making it central to , differential equations, and . Additionally, exp(x) satisfies key functional equations, such as exp(x + y) = exp(x) · exp(y) and exp() = , which underscore its role in modeling multiplicative processes. The is ubiquitous in science and , describing phenomena like , , , and electrical circuits, where rates of change are proportional to the quantity involved. In more advanced contexts, it extends to the as exp(z) = e^z for complex z, forming the basis for e^{iθ} = cos(θ) + i sin(θ), which links exponential and . Its importance stems from these properties, positioning exp(x) as a cornerstone of mathematical modeling across disciplines.

Definitions

Differential equation approach

One historical motivation for the exponential function arises from the modeling of continuous compounding interest, as investigated by Jacob Bernoulli in 1683. Bernoulli considered the growth of an investment where interest is compounded infinitely often, leading to a scenario where the instantaneous rate of change of the principal is proportional to the principal itself. This proportionality results in the differential equation \frac{dP}{dt} = r P(t), where P(t) is the value at time t and r is the interest rate. To solve this equation, apply separation of variables: \frac{dP}{P} = r \, dt. Integrating both sides yields \ln |P| = r t + C, so P(t) = K e^{r t} for some constant K > 0. With initial condition P(0) = P_0, it follows that K = P_0, giving P(t) = P_0 e^{r t}. For the standard exponential function, set r = 1 and P_0 = 1, yielding the initial value problem f'(x) = f(x) with f(0) = 1, whose solution is f(x) = e^x. The uniqueness of this solution follows from the , which guarantees a unique local solution to the y' = g(x, y), y(x_0) = y_0, when g is continuous and continuous in y. Here, g(x, y) = y satisfies these conditions globally, ensuring the solution e^x is unique on \mathbb{R}. The e is defined as the value f(1), representing the unique factor by which the solution grows over the interval [0, 1]. This characterization aligns with the property that the equals the function itself, distinguishing e \approx 2.71828 as the for which the defining the at zero equals 1.

Power series definition

The \exp(x) can be defined for all real numbers x by its expansion around zero: \exp(x) = \sum_{n=0}^{\infty} \frac{x^n}{n!}. This representation arises as the of \exp(x), where the coefficients are determined by repeated : assuming \exp(x) is infinitely differentiable with \exp^{(n)}(x) = \exp(x) for all n \geq 0 and \exp(0) = 1, the Taylor coefficients at zero are \exp^{(n)}(0)/n! = 1/n!. The series converges for every real x, as established by the ratio test applied to the absolute values of consecutive terms: \lim_{n \to \infty} \left| \frac{x^{n+1}/(n+1)!}{x^n/n!} \right| = \lim_{n \to \infty} \frac{|x|}{n+1} = 0 < 1. This infinite radius of convergence reflects the rapid growth of the factorial denominator n!, which outpaces any polynomial growth in the numerator. Evaluating the series at x = 1 yields the base of the natural logarithm, e = \exp(1) = \sum_{n=0}^{\infty} \frac{1}{n!}, a fundamental constant approximately equal to 2.71828 whose partial sums provide increasingly accurate approximations due to the series' convergence properties. Term-by-term differentiation of the power series reproduces the original series, confirming that \exp(x) satisfies the differential equation y' = y with initial condition y(0) = 1. Since the power series converges absolutely for every complex number z, it defines the analytic continuation of \exp(x) to the entire complex plane, rendering \exp(z) an entire function with no singularities.

Limit of powers definition

The exponential function \exp(x) for real x is defined as the limit \exp(x) = \lim_{n \to \infty} \left(1 + \frac{x}{n}\right)^n, where n is a positive integer. To establish the existence of this limit, consider the case x > 0. The binomial theorem expands \left(1 + \frac{x}{n}\right)^n = \sum_{k=0}^n \binom{n}{k} \left(\frac{x}{n}\right)^k, revealing that the sequence a_n = \left(1 + \frac{x}{n}\right)^n is increasing because each term in the expansion for a_{n+1} exceeds the corresponding term in a_n. Furthermore, the sequence is bounded above, as the expansion shows a_n < \sum_{k=0}^\infty \frac{x^k}{k!} < \infty by comparing to a geometric series after bounding the coefficients. Thus, by the monotone convergence theorem, the limit exists and is finite. For x = 0, the limit is trivially 1, and for x < 0, the limit exists by considering the reciprocal. In the specific case x = 1, the constant e is defined as e = \lim_{n \to \infty} \left(1 + \frac{1}{n}\right)^n \approx 2.71828. Numerical approximations illustrate the convergence: for n = 10, the value is about 2.59374; for n = 100, about 2.70481; for n = 1000, about 2.71692; and for n = 10^6, about 2.718280. This limit definition extends to powers via manipulation: \exp(xy) = \lim_{n \to \infty} \left(1 + \frac{xy}{n}\right)^n = \lim_{n \to \infty} \left[ \left(1 + \frac{x}{n}\right)^n \right]^y = [\exp(x)]^y for real y. Historically, this formulation ties to 's 1683 investigation of compound interest, where continuous compounding led to the limit \lim_{n \to \infty} \left(1 + \frac{1}{n}\right)^n as the effective growth factor.

Inverse logarithm definition

The natural logarithm function, denoted \ln x, is defined for x > 0 as the definite integral \ln x = \int_1^x \frac{1}{t} \, dt. This integral representation establishes \ln x as a strictly increasing and continuous function, with derivative \frac{d}{dx} \ln x = \frac{1}{x} > 0 for x > 0, ensuring it is bijective from (0, \infty) onto (-\infty, \infty). Consequently, \ln x possesses a unique , called the and denoted \exp y or e^y, defined on all real numbers y \in \mathbb{R} with (0, \infty), satisfying \exp(\ln x) = x for all x > 0 and \ln(\exp y) = y for all real y. This inverse relation guarantees that \exp y > 0 for every real y, as the of \ln x covers all reals while its domain is the positive reals. The uniqueness of \exp y as the inverse stems directly from the monotonicity and continuity of \ln x, which ensure a one-to-one correspondence without ambiguity in the inversion. This definition highlights the mutual exclusivity of the exponential and natural logarithm in real analysis, where the exponential maps reals to positives and the logarithm reverses this mapping. For a general base b > 0, b \neq 1, the exponential function b^x is expressed via the change of base as b^x = \exp(x \ln b), allowing extension from the natural base while preserving the core inverse properties. As a consequence of this inverse relationship, the exponential function satisfies the differential equation y' = y with initial condition y(0) = 1.

Fundamental Properties

Algebraic properties

The satisfies the fundamental \exp(x + y) = \exp(x) \exp(y) for all real numbers x and y. This establishes the as a from the additive group of real numbers to the of . A proof follows from the power series representation \exp(z) = \sum_{n=0}^\infty \frac{z^n}{n!}, which converges for all z. Multiplying the series for \exp(x) and \exp(y) yields the \sum_{n=0}^\infty \sum_{k=0}^n \frac{x^{n-k} y^k}{(n-k)! k!}, and applying the gives \sum_{n=0}^\infty \frac{(x+y)^n}{n!} = \exp(x+y). Setting x = [0](/page/0) in the implies \exp([0](/page/0)) = \exp([0](/page/0)) \exp([0](/page/0)), so \exp([0](/page/0)) (\exp([0](/page/0)) - 1) = [0](/page/0). Since the power series at z = [0](/page/0) evaluates to 1, it follows that \exp([0](/page/0)) = 1 > [0](/page/0). Substituting y = -x yields \exp(x) \exp(-x) = \exp([0](/page/0)) = 1, so \exp(-x) = 1 / \exp(x). The is strictly positive for all real x, as the power series terms are all nonnegative for x \geq [0](/page/0) (hence \exp(x) \geq 1 > [0](/page/0)), \exp([0](/page/0)) = 1 > [0](/page/0), and \exp(x) = 1 / \exp(-x) > [0](/page/0) for x < [0](/page/0). The functional equation extends to integer multiples: \exp(n x) = [\exp(x)]^n for any integer n. For positive integers, this holds by induction, as \exp((n+1)x) = \exp(n x + x) = \exp(n x) \exp(x) = [\exp(x)]^{n+1}, with the base case \exp(x) = [\exp(x)]^1. The case n = 0 follows from \exp(0) = 1 = [\exp(x)]^0. For negative integers, use \exp(-n x) = 1 / \exp(n x) = [\exp(x)]^{-n}. This identity generalizes to rational exponents: for r = p/q with integers p, q (q > 0), let y = \exp(x/q), then y^q = \exp(x), so y = [\exp(x)]^{1/q} (the unique positive real q-th root, as \exp > 0), and \exp(r x) = \exp(p \cdot (x/q)) = [ \exp(x/q) ]^p = [\exp(x)]^r. Beyond these identities and the power series expansion, the exponential function has no closed-form expression in terms of algebraic functions; it is transcendental, meaning it cannot be expressed as a finite combination of polynomials and rational functions using the field operations of , , , and division.

Analytic properties

The exponential function \exp(x), often denoted e^x, satisfies the \frac{d}{dx} \exp(x) = \exp(x), making it equal to its own . This property follows from the \exp(x + y) = \exp(x) \exp(y), which underpins the limit definition of the . Consequently, all higher-order derivatives of \exp(x) are also \exp(x), as repeated differentiation yields the same result each time. The indefinite integral of the is given by \int \exp(x) \, dx = \exp(x) + C, where C is the constant of integration, mirroring its behavior. For real arguments, \exp(x) is strictly increasing because its \exp(x) > 0 for all x, ensuring that \exp(x_1) < \exp(x_2) whenever x_1 < x_2. This monotonicity implies that \exp(x) is bijective from \mathbb{R} to (0, \infty), with its inverse being the natural logarithm function \ln(y). In the complex plane, the exponential function extends to \exp(z) for z \in \mathbb{C}, which is an entire function—holomorphic everywhere with no singularities.

Growth and behavior

The exponential function e^x exhibits distinct asymptotic behaviors at the extremes of its domain. As x \to \infty, e^x \to \infty, demonstrating unbounded growth. Conversely, as x \to -\infty, e^x \to 0, approaching the x-axis without reaching it. The graph of y = e^x is a smooth curve that passes through the point (0, 1) and lies entirely above the x-axis for all real x. It is strictly increasing everywhere, as confirmed by its analytic derivative being positive, and convex, meaning the curve bends upward and any line segment connecting two points on the graph lies above the curve. At the origin, the graph is tangent to the line y = x + 1. Compared to polynomial functions, e^x grows faster than any polynomial of finite degree as x \to \infty; for example, \lim_{x \to \infty} \frac{e^x}{x^n} = \infty for any fixed n. This rapid growth underscores its superpolynomial nature. While aperiodic on the real numbers—lacking any real period that repeats its values—the complex exponential function e^z is periodic with period $2\pi i, leading to oscillatory behavior in the imaginary direction.

Generalizations and Extensions

Exponential functions with other bases

The exponential function with an arbitrary base b > 0, b \neq 1, is defined for all real exponents x by
b^x = \exp(x \ln b),
where \exp denotes the and \ln is the natural logarithm. This definition leverages the natural exponential as the foundational form while extending its structure to other positive bases excluding 1, ensuring the function remains positive and well-defined over the reals.
Key algebraic properties hold under this definition, mirroring those of the natural exponential. For instance, the addition formula
b^{x+y} = b^x b^y
applies for all real x and y, derived from the property \exp(u + v) = \exp(u) \exp(v) with u = x \ln b and v = y \ln b. Similarly, b^{xy} = (b^x)^y and b^{-x} = 1 / b^x follow analogously.
The base change formula relates exponentials across different bases using logarithms. Specifically, for bases a > 0, a \neq 1, and b > 0, b \neq 1,
b^x = a^{x \log_a b},
where \log_a b = \ln b / \ln a. To prove this equivalence, take the natural logarithm of both sides of the proposed identity: \ln(b^x) = x \ln b. Substituting the right side yields \ln(a^{x \log_a b}) = x \log_a b \cdot \ln a = x (\ln b / \ln a) \ln a = x \ln b, confirming the equality since the natural exponential is .
Special cases illustrate practical uses. When b = 10, the function $10^x is the of the \log_{10}, employed in decimal-based scientific calculations and engineering contexts. For b = 2, $2^x models processes, such as via in , where each doubles the . These functions inherit analytic from their . Since \exp and \ln are continuous and infinitely differentiable on the reals, the composition b^x is continuous for all real x and differentiable with derivative
\frac{d}{dx} b^x = b^x \ln b.
This holds everywhere, with the factor \ln b determining growth rate (positive if b > 1, negative if $0 < b < 1).

Complex exponential function

The complex exponential function extends the real exponential to complex arguments, providing a unified framework for trigonometric and hyperbolic functions. For a complex number z = x + iy with real parts x and y, it is defined as \exp(z) = \exp(x + iy) = \exp(x) (\cos y + i \sin y), where \exp(x) is the standard real exponential function. This decomposition arises from the entire nature of the exponential as an analytic function in the complex plane, with the power series definition converging everywhere. The relation \exp(iy) = \cos y + i \sin y is known as Euler's formula, establishing a deep connection between exponential growth and rotational motion in the complex plane. This formula, derived from solving the differential equation \dot{z} = i z with initial condition z(0) = 1, parametrizes the unit circle as y varies over the reals. A key property of the complex exponential is its periodicity: \exp(z + 2\pi i) = \exp(z) for any complex z, with $2\pi i as the fundamental period. This periodicity reflects the $2\pi-periodicity of the cosine and sine functions in the imaginary part, causing the function to repeat values along vertical lines parallel to the imaginary axis in the complex plane. The trigonometric functions can be expressed directly in terms of the complex exponential: the real part gives \cos y = \operatorname{Re}(\exp(i y)), and the imaginary part yields \sin y = \operatorname{Im}(\exp(i y)). More explicitly, \cos y = \frac{\exp(i y) + \exp(-i y)}{2}, \quad \sin y = \frac{\exp(i y) - \exp(-i y)}{2i}. These identities facilitate the analysis of oscillatory phenomena through exponential forms. As the inverse of the exponential function, the complex logarithm is multi-valued due to the periodicity of \exp(z). For a nonzero complex number w, the solutions to \exp(z) = w are z = \log|w| + i (\arg w + 2\pi n) for integers n, requiring branch cuts—typically along the negative real axis—to define a single-valued principal branch. The origin serves as a branch point, where the function cannot be continuously defined without encircling it leading to different values.

Matrix and operator exponentials

The matrix exponential of a square matrix A, denoted \exp(A) or e^A, is defined by the power series \exp(A) = \sum_{n=0}^{\infty} \frac{A^n}{n!}, where A^0 = I is the identity matrix and A^n denotes the n-th matrix power for n \geq 1. This series converges absolutely to a well-defined matrix for every square matrix A over the real or complex numbers, regardless of the eigenvalues of A. Key properties of the matrix exponential mirror those of the scalar exponential when applicable. If matrices A and B commute (i.e., AB = BA), then \exp(A + B) = \exp(A) \exp(B) = \exp(B) \exp(A). More generally, for any scalar t, the matrix \exp(tA) provides the fundamental solution to the linear system of ordinary differential equations X' = AX with initial condition X(0) = I, where X(t) is a matrix-valued function. The matrix exponential interacts naturally with the spectral properties of A. If v is an eigenvector of A with eigenvalue \lambda (i.e., Av = \lambda v), then v is also an eigenvector of \exp(A) with eigenvalue e^\lambda. This relation extends to the case where A is diagonalizable, allowing \exp(A) to be computed via the eigenvalues of A. When A is a $1 \times 1 matrix, the matrix exponential reduces to the standard scalar exponential function. The concept generalizes to linear operators on finite-dimensional vector spaces, where the exponential is defined analogously via the power series, with convergence guaranteed in the operator norm. This framework underpins applications in solving linear dynamical systems, such as those arising in control theory and physics.

Applications

In differential equations and calculus

The exponential function serves as the fundamental solution to linear homogeneous first-order ordinary differential equations of the form y' - k y = 0, where k is a constant. The general solution is y(x) = C \exp(k x), with C an arbitrary constant, reflecting the exponential growth or decay depending on the sign of k. This form arises directly from separation of variables or recognition of the derivative of the exponential, establishing \exp(k x) as the eigenfunction for the constant-coefficient operator \frac{d}{dx} - k. For nonhomogeneous first-order linear differential equations y' + p(x) y = q(x), the integrating factor method leverages the exponential function to simplify the equation. The integrating factor is defined as \mu(x) = \exp\left( \int p(x) \, dx \right), which, when multiplied through the equation, yields \frac{d}{dx} \left[ y \mu(x) \right] = q(x) \mu(x). Integrating both sides then provides the solution y(x) = \frac{1}{\mu(x)} \left( \int q(x) \mu(x) \, dx + C \right), highlighting the exponential's role in rendering the left-hand side exact. This technique, originally developed by , transforms variable-coefficient problems into integrable forms without altering the underlying structure. In the context of Laplace transforms, the exponential function appears as the kernel that facilitates solving initial-value problems for linear differential equations. The unilateral Laplace transform is given by \mathcal{L}\{ f(t) \}(s) = \int_0^\infty f(t) e^{-s t} \, dt, where e^{-s t} weights the integral to ensure convergence for functions of exponential order, converting differentiation into multiplication by s in the s-domain. This property allows straightforward algebraic manipulation of transformed differential equations, with the inverse transform recovering time-domain solutions via exponential-based residues or tables. The method of variation of parameters for nonhomogeneous linear differential equations with constant coefficients relies on an exponential basis for the homogeneous solutions. For a second-order equation y'' + a y' + b y = g(x), the fundamental solutions are typically y_1(x) = \exp(r_1 x) and y_2(x) = \exp(r_2 x) (or modified for repeated roots), where r_1, r_2 are roots of the characteristic equation. A particular solution is then sought as y_p(x) = u_1(x) y_1(x) + u_2(x) y_2(x), with u_1' and u_2' determined by solving the system u_1' y_1 + u_2' y_2 = 0, \quad u_1' y_1' + u_2' y_2' = g(x), yielding integrals involving the Wronskian, which for exponential bases simplifies to \exp((r_1 + r_2) x). This approach generalizes to higher orders, emphasizing the exponential's centrality in parametric variation. For systems of linear differential equations, the matrix exponential provides an analogous solution framework.

In modeling natural phenomena

The exponential function models continuous compounding in finance, where interest is added instantaneously rather than at discrete intervals. The amount A(t) after time t for a principal P at rate r is given by A(t) = P \exp(rt), derived as the limit \lim_{n \to \infty} P \left(1 + \frac{r}{n}\right)^{nt} when compounding frequency n approaches infinity. This formulation captures uninterrupted growth, essential for precise long-term projections in investments and economics. In nuclear physics, the exponential function describes radioactive decay, where the number of undecayed nuclei N(t) at time t follows N(t) = N_0 \exp(-\lambda t), with N_0 as the initial number and \lambda the decay constant. This law, first formulated by Ernest Rutherford and Frederick Soddy, reflects the probabilistic nature of atomic disintegration, where the decay rate is proportional to the current number of nuclei. The half-life T_{1/2}, the time for half the nuclei to decay, relates to \lambda via \lambda = \frac{\ln 2}{T_{1/2}}, enabling predictions of material stability in fields like medicine and geology./University_Physics_III_-Optics_and_Modern_Physics(OpenStax)/10%3A__Nuclear_Physics/10.04%3A_Radioactive_Decay) Exponential growth also applies to biological populations under ideal conditions, modeled by the differential equation \frac{dP}{dt} = k P, yielding P(t) = P_0 \exp(kt) where P_0 is the initial population and k > 0 the growth rate. This idea traces to Thomas Malthus, who argued that populations tend to increase geometrically when unchecked by resources, influencing and . Examples include early-stage bacterial cultures or unchecked species in new habitats, where reproduction rates amplify rapidly. However, pure models overlook environmental constraints, leading to unrealistic unbounded . The logistic model corrects this by incorporating a K, modifying the growth rate to \frac{dP}{dt} = k P \left(1 - \frac{P}{K}\right), which slows as P approaches K due to resource limits like food or space. Introduced by Pierre-François Verhulst, this adjustment better fits observed patterns in , such as populations stabilizing after initial booms.

In probability and statistics

In probability and statistics, the exponential function plays a central role in defining key distributions and processes, particularly those modeling waiting times and event occurrences under constant rates. The , denoted as Exp(λ) where λ > 0 is the rate parameter, describes the time between events in a process with a constant probability rate of occurrence. Its is given by f(x) = \lambda e^{-\lambda x}, \quad x \geq 0, and the cumulative distribution function is F(x) = 1 - e^{-\lambda x} for x \geq 0. The of the is $1/\lambda, representing the expected waiting time, while the variance is $1/\lambda^2. A defining feature is its memoryless property: for any s, t \geq 0, P(X > s + t \mid X > s) = P(X > t) = e^{-\lambda t}, meaning the distribution of remaining time is of elapsed time, which distinguishes it from other waiting time models. The arises naturally in the process, a fundamental model for counting random events over time or space at a constant average rate λ. In a process, the interarrival times—the durations between successive events—are and exponentially distributed with rate λ. This connection implies that the number of events in a fixed interval of length t follows a with parameter λt, enabling the modeling of phenomena like particle arrivals or customer inflows where events occur independently and at a steady rate. The (MGF) of an X ~ Exp(λ) further highlights the exponential function's utility in deriving statistical properties. The MGF is defined as M_X(t) = E[e^{tX}] = \frac{\lambda}{\lambda - t} for t < \lambda, which facilitates computation of : the first () is the at t=0, yielding 1/λ, and higher follow similarly. This form also aids in analyzing sums of exponentials, which relate to gamma distributions in more complex models. Applications of the exponential distribution extend to survival analysis, where it models lifetimes or durations under constant hazard rates, such as in reliability engineering for failure times. In this context, the survival function S(t) = e^{-\lambda t} represents the probability of surviving beyond time t, with the constant hazard h(t) = \lambda implying no aging effect, ideal for scenarios like electronic component failures. In queueing theory, exponential interarrival and service times underpin Markovian models like the M/M/1 queue, where steady-state probabilities depend on the traffic intensity ρ = λ/μ (with μ the service rate), enabling analysis of wait times and system utilization in service systems such as call centers.

History and Computation

Historical development

The development of the exponential function and the constant e traces its roots to early efforts in understanding logarithms and growth processes in the . In 1614, Scottish mathematician published Mirifici Logarithmorum Canonis Descriptio, introducing the concept of logarithms as a tool to simplify complex calculations, particularly in astronomy and . These logarithms were not exactly natural logarithms but were proportional to them, serving as a precursor to the inverse relationship that would later define the exponential function as the inverse of the natural logarithm. Napier's work laid foundational groundwork by establishing a that facilitated through , implicitly hinting at the nature of antilogarithms. A significant advancement came in 1683 when Jacob Bernoulli, while investigating continuous compound interest, encountered the constant e through the limit of (1 + 1/n)^n as n approaches infinity. Bernoulli's analysis in his work on interest accumulation revealed that this limit converges to approximately 2.71828, marking the first explicit recognition of e as a distinct mathematical constant arising from iterative growth processes. This discovery highlighted e's role in modeling unbounded growth, such as in financial applications, without yet connecting it to a broader exponential framework. The formalization of the occurred in the through the contributions of Leonhard Euler. In his seminal 1748 treatise , Euler introduced the notation e^x for the exponential function and defined it via its expansion, solidifying e as the base of the natural logarithm. Euler's work integrated the constant into , demonstrating its properties through infinite series and differential equations, and established the exponential as a fundamental function in . This publication represented a pivotal moment, transforming Bernoulli's isolated observation into a cornerstone of mathematical theory. The 19th century brought rigorous proofs concerning the nature of , elevating its status in . In 1873, Charles Hermite provided the first proof that is a , meaning it is not the root of any non-zero equation with rational coefficients. Hermite's demonstration, using integral representations and properties of entire functions, resolved a long-standing and underscored 's beyond mere . Building on this, extended the result in 1882 by proving the transcendence of π, which relied on Hermite's methods and further affirmed the exponential function's deep connections to . These proofs marked a high point in the historical evolution, confirming the exponential's transcendence and its implications for and .

Numerical computation methods

One common method for numerically evaluating the \exp(x) involves truncating its expansion \sum_{k=0}^{\infty} \frac{x^k}{k!} for small |x|, typically |x| < 1, where is rapid and few terms suffice for double-precision accuracy. The approximation is \exp(x) \approx \sum_{k=0}^{n} \frac{x^k}{k!}, with the bounded by the Lagrange |R_n(x)| \leq \frac{|x|^{n+1}}{(n+1)!} e^{|x|}, allowing reliable error estimation and choice of n based on desired precision. This approach is widely implemented in mathematical libraries, as the denominators grow quickly, minimizing computational cost for n \approx 10-15 terms in . For values of x near zero, direct evaluation of \exp(x) - 1 incurs significant loss of precision due to subtractive cancellation, as \exp(x) \approx 1 + x and the leading 1 dominates in floating-point representation. The specialized function \operatorname{expm1}(x) = \exp(x) - 1 addresses this by employing tailored algorithms, such as argument reduction to small ranges followed by series summation or table-driven evaluation, achieving full machine precision even for |x| < 10^{-8}. These implementations, often using IEEE 754-compliant table methods, are essential in applications like where small differences matter. Continued fraction expansions offer an alternative for rational approximations of \exp(x), particularly in intervals like [- \ln 2, \ln 2], where they converge faster than truncated series for certain error tolerances. The Gaussian continued fraction form yields near-minimax rational approximants that outperform simple convergents, with explicit coefficients enabling efficient recursive computation and error bounds below $10^{-10} using modest-degree fractions. This is advantageous for software libraries requiring balanced accuracy across positive and negative arguments, as the reciprocal property simplifies handling. In settings, such as FPGAs or systems, lookups accelerate \exp(x) by precomputing values over input ranges and interpolating via low-order polynomials, reducing to a few clock cycles while maintaining 16-32 bit precision. The (COordinate Rotation DIgital Computer) provides a multiplier-free alternative, computing \exp(x) through iterative shifts and adds in mode, ideal for resource-constrained devices despite slower requiring 20-30 iterations for high accuracy. These techniques are often hybridized with series for fractional parts to optimize power and area in applications like .

References

  1. [1]
    Calculus I - Exponential Functions - Pauls Online Math Notes
    Nov 16, 2022 · An exponential function is f(x)=bx, where b>0 and b≠1. It always takes the value of 1 at x=0 and is always positive.
  2. [2]
    9.3 The exponential function
    The natural exponential function, exp(x), is the inverse of ln(x), with domain all real numbers and range (0,∞). It's often written as ex.
  3. [3]
    The exponential function - Math Insight
    An exponential function has the independent variable as the exponent, and it can describe growth or decay, like f(x)=2x.
  4. [4]
    [PDF] 1 Definition and Properties of the Exp Function
    The exp function E(x) = ex is the inverse of the log function. L(x) = ln x: L ◦ E(x) = ln ex = x, ∀x. Properties. • ln x is the inverse of ex: ∀x > 0, E ◦ L = ...
  5. [5]
    Exponential Functions - Ximera - The Ohio State University
    We say that a function is exponential whenever its algebraic form is for some positive constants and where . (Technically, the formal definition of an ...
  6. [6]
    [PDF] Exponential Functions
    An exponential function is defined as f(x) = ax, where 'a' is the base, which must be a positive number. The variable 'x' is the exponent.
  7. [7]
    Definition and Examples - BioMath: Exponential Functions
    Exponential functions can be used to understand many natural phenomena, including bacterial growth and radioactive decay.
  8. [8]
    Exponential Functions - Department of Mathematics at UTSA
    Oct 21, 2021 · An exponential function is a function where a constant base (b) is raised to a variable.
  9. [9]
    Tutorial 42: Exponential Functions. - West Texas A&M University
    Mar 21, 2011 · 1, and the exponent x is any real number, is called an exponential function. Again, note that the variable x is in the exponent as opposed to ...<|control11|><|separator|>
  10. [10]
    The number e - MacTutor History of Mathematics
    It may have been Jacob Bernoulli who first understood the way that the log function is the inverse of the exponential function. On the other hand the first ...<|separator|>
  11. [11]
  12. [12]
    [PDF] 1. THE EXPONENTIAL FUNCTION. The solution of first-order linear ...
    THE EXPONENTIAL FUNCTION. The solution of first-order linear differential equations is based on the exponential, so it is useful to recall its definition ...
  13. [13]
    [PDF] Picard's Existence and Uniqueness Theorem
    These notes on the proof of Picard's Theorem follow the text Fundamentals of Differential. Equations and Boundary Value Problems, 3rd edition, by Nagle, ...<|separator|>
  14. [14]
    Exploring the derivative of the exponential function - Math Insight
    A guided tour into the reasons that the derivative of the exponential function with base e is the function itself.
  15. [15]
    DLMF: §4.2 Definitions ‣ Logarithm, Exponential, Powers ‣ Chapter ...
    §4.2(iii) The Exponential Function. ⓘ. Keywords: analytic properties, definition, exponential function, periodicity, power series, zeros; Notes: See Hobson ...
  16. [16]
    [PDF] Power Series - UC Davis Math
    The ratio test gives a simple, but useful, way to compute the radius of convergence, although it doesn't apply to every power series. Theorem 6.4. Suppose that ...
  17. [17]
    The Number e and the Exponential Function - Galileo
    The Exponential Function e​​ Taking our definition of e as the infinite n limit of (1+1n)n, it is clear that ex is the infinite n limit of (1+1n)nx. Let us write ...
  18. [18]
    [PDF] The Exponential Function
    for all n ∈ N. Thus ex = lim n→∞ (1 + x n)n. ≥ 1 + x. To prove part (b), suppose that x < 1. Then 1 − x > 0 and by part (a) e−x ≥ 1 − x > 0. Rearranging, we ...
  19. [19]
    convergence of the sequence (1+1/n)^n - PlanetMath.org
    Mar 22, 2013 · In order to prove part 1, consider the binomial expansion for an a n : an=n∑k=0(nk)1nk=n∑k=01k!nnn−1n…n−(k−1)n=n∑k=01k!(1−1n)…
  20. [20]
  21. [21]
    9.2 The natural logarithm
    Definition 9.2.1 The natural logarithm ln(x) is an antiderivative of 1/x, given by lnx=∫x11tdt. ◻. Figure 9.2.1 gives a geometric interpretation of ln.Missing: dt/ | Show results with:dt/
  22. [22]
    [PDF] The Natural Logarithm - Math
    ln(x) := ∫ x. 1. 1 t dt with domain (0, ∞). Let's compare this function with some others: ∫ x. 1. 1dt = x − 1 (since t is an antiderivative of 1). ∫ x. 1 tdt =.
  23. [23]
    [PDF] The exponential function (Sect. 7.3) The inverse of the logarithm
    The exponential function, exp : R → (0,∞), is the inverse of the natural logarithm, that is, exp(x) = y ⇔ x = ln(y). e y.
  24. [24]
    [PDF] Section 6.3, The Natural Exponential Function
    The inverse of ln is called the natural exponential function and is denoted exp. Therefore x = exp y ⇔ y = ln x. Note: exp x = ex where e is the unique real ...
  25. [25]
    [PDF] Exponential and logarithm functions
    Fact 10. For any base b, bx = ex ln b. Proof: The natural log of bx is ln(bx) = xlnb, and this means that bx = eln(bx) = ex ln b. Comment: These two facts will ...
  26. [26]
    [PDF] Chapter 10: Power Series - UC Davis Math
    First, we show that exey = ex+y. For the moment, we continue to write the function ex as E(x) to emphasise that we use nothing beyond its power series.
  27. [27]
    DLMF: §4.8 Identities ‣ Logarithm, Exponential, Powers ‣ Chapter ...
    This is interpreted that every value of Ln ( z 1 ⁢ z 2 ) is one of the values of Ln ⁡ z 1 + Ln ⁡ z 2 , and vice versa.
  28. [28]
    [PDF] Transcendental Functions
    The exponential function, like the trigonometric functions, is a transcendental function. These are func- tions which cannot be expressed as a quotient of ...
  29. [29]
    Exponential Function -- from Wolfram MathWorld
    The most general form of an exponential function is a power-law function of the form f(x)=ab^(cx+d), where a, c, and d are real numbers, b is a positive real ...
  30. [30]
    [PDF] ON THE EXPONENTIAL FUNCTION 1. Definitions of the exponential ...
    We can choose any definition from the above, and as soon as we choose one, the others become properties of the function. So the ques-.
  31. [31]
    [PDF] Lesson 15. Exponential function exp(z) = e (cosy + ... - Purdue Math
    Exponential function exp(z) = e x. (cosy + isiny), u = e x cosy, v = e x ... A function analytic in C is called entire, thus exp(z) is an entire function.
  32. [32]
    Exponential Functions - Definition, Formula, Asymptotic Behaviour ...
    Asymptotic Behaviour of Exponential Functions​​ Asymptotic behaviour refers to how a function behaves as its input values approach a certain boundary, either ...
  33. [33]
    4.1 - Exponential Functions and Their Graphs
    Transcendental functions return values which may not be expressible as rational numbers or roots of rational numbers.
  34. [34]
    Exponential Functions - BYJU'S
    Exponential Function Properties · The graph passes through the point (0,1). · The domain is all real numbers · The range is y>0 · The graph is increasing · The graph ...
  35. [35]
    [PDF] Convexity
    The logarithm f(x) = log x is concave on the interval 0 <x< ∞, and the exponential f(x) = ex is convex everywhere. 3. f(x) is convex iff −f(x) is concave.
  36. [36]
    6. Derivative of the Exponential Function - Interactive Mathematics
    Since the derivative of ex is ex, then the slope of the tangent line at x = 2 is also e2 ≈ 7.39. We can see that it is true on the graph: The graph of ...<|control11|><|separator|>
  37. [37]
    [PDF] relative rates of growth - UNM Math
    Ex 3: You may have noticed that exponential functions like 2x and ex seem to grow more rapidly as x gets large than polynomials and rational functions. Figure 3 ...
  38. [38]
    periodicity of exponential function - PlanetMath
    Feb 20, 2014 · Proof. Let ω ω be any period of the exponential function Dlmf Mathworld Planetmath , i.e. ez+ω=ezeω=ez e z + ω = e z ⁢ e ω = e z for all z∈C ...
  39. [39]
    [PDF] 3 Exponential Functions, Logarithms, and e
    b. )x . Logarithms Base 2. Definition 2 The exponential function with base b is of the form f(x) = bx, where b > 0 and b 6= 1. With a positive base b, raised ...
  40. [40]
    Algebra - Logarithm Functions - Pauls Online Math Notes
    Nov 16, 2022 · So, the common logarithm is simply the log base 10, except we drop the “base 10” part of the notation. Similarly, the natural logarithm is ...
  41. [41]
    t - BioMath: Exponential Functions
    Bacterial growth by binary fission, for instance, can be easily described using a base 2 exponential function, because each round of cell division doubles the ...
  42. [42]
    Calculus I - Derivatives of Exponential and Logarithm Functions
    Nov 16, 2022 · In this section we derive the formulas for the derivatives of the exponential and logarithm functions.
  43. [43]
    [PDF] 6. The complex exponential
    Euler's formula: (4) eit = cost + isint. In this formula, the left hand side is by definition the solution to ˙z = iz such that z(0) = 1. The right hand side ...Missing: authoritative sources
  44. [44]
  45. [45]
    [PDF] Matrix-Exponentials | MIT
    Sep 7, 2017 · 3 Series definition of a matrix exponential. Just plugging in t = 1 above, we see that we have defined the matrix exponential by. eA = XeΛX−1.
  46. [46]
    The Exponential of a Matrix
    The matrix exponential e^At is defined as a power series: e^At = sum from n=0 to infinity of t^n A^n / n!
  47. [47]
    [PDF] 4. Linear Systems: Matrix Exponentials - UMD MATH
    Aug 2, 2022 · The matrix exponential of A, denoted as etA or exp(tA), is the solution of the matrix-valued initial-value problem (4.2), and is not the matrix ...
  48. [48]
    [PDF] Notes on the Matrix Exponential and Logarithm Howard E. Haber
    Nov 14, 2023 · These notes summarize important properties of the matrix exponential and logarithm, including a series expansion for dlnA(t)/dt.
  49. [49]
    [PDF] The exponential function for matrices - UCR Math Department
    Matrix exponentials provide a concise way of describing the solutions to systems of homoge- neous linear differential equations that parallels the use of ...
  50. [50]
    [PDF] Differential Equations
    The equation ˙y = ky, or ˙y - ky = 0 is linear and homogeneous, with a particularly simple p(t) = -k. ln |y| = P(t) + C y = 소eP (t)+C y = AeP (t),
  51. [51]
    [PDF] FIRST ORDER DIFFERENTIAL EQUATIONS - UCSD Math
    1. Solve the corresponding homogeneous equation y0 + ky = 0. The solution is yh = Ce−kt. This is where the arbitrary constant appears!
  52. [52]
    [PDF] 1.5 Linear First-Order Equations
    Begin by calculating the integrating factor p(x) = e√ P(x) dx. 2. Then multiply both sides of the differential equation by p(x). 3. Next, recognize the left ...Missing: ∫ | Show results with:∫
  53. [53]
    [PDF] Solving Linear First-Order Differential Equations Leonard Euler's ...
    For more information, please contact aprock@ursinus.edu. Page 2. Solving Linear First-Order Differential Equations. Leonard Euler's Integrating Factor Method.Missing: exp( | Show results with:exp(
  54. [54]
    Differential Equations - Laplace Transforms - Pauls Online Math Notes
    Nov 16, 2022 · In this section we introduce the way we usually compute Laplace transforms that avoids needing to use the definition.
  55. [55]
    The Laplace Transform of Functions
    To productively use the Laplace Transform, we need to be able to transform functions from the time domain to the Laplace domain. We can do this by applying the ...
  56. [56]
    Differential Equations - Variation of Parameters
    Nov 16, 2022 · In this section we introduce the method of variation of parameters to find particular solutions to nonhomogeneous differential equation.Missing: basis | Show results with:basis
  57. [57]
    Matrix Exponentials | Differential Equations - MIT OpenCourseWare
    This section provides materials for a session on the basic linear theory for systems, the fundamental matrix, and matrix-vector algebra.
  58. [58]
    [PDF] Calculation of Continuous Compound Interest - eGrove
    Substituting this value in the first equation above, the interest formula for continuous compounding is found to be a = p ert. We now have our exact ...
  59. [59]
    The Rule of 72 - Stanford University
    Continuous compounding. For continuous compounding the derivation is simpler: implies. or. Using 100r to get percentages and taking 70 as a close enough ...
  60. [60]
    Radioactive Decay Law (Rutherford–Soddy) | Request PDF
    The formulation of the radioactive decay law, in 1902, by Ernest Rutherford (1871–1937) and Frederick Soddy (1877–1956) was part of a number of discoveries ...
  61. [61]
    An Essay on the Principle of Population [1798, 1st ed.]
    In this work Malthus argues that there is a disparity between the rate of growth of population (which increases geometrically) and the rate of growth of ...Missing: exponential | Show results with:exponential
  62. [62]
    Verhulst and the logistic equation (1838) - ResearchGate
    In 1838 the Belgian mathematician Verhulst introduced the logistic equation, which is a kind of generalization of the equation for exponential growth.
  63. [63]
    1.3.6.6.7. Exponential Distribution - Information Technology Laboratory
    The exponential distribution is primarily used in reliability applications. The exponential distribution is used to model data with a constant failure rate ...Missing: engineering | Show results with:engineering
  64. [64]
    [PDF] 18.440: Lecture 20 Exponential random variables
    ▷ Suppose X is exponential with parameter λ. ▷ Memoryless property: If X represents the time until an event occurs, then given that we have seen no event up to.
  65. [65]
    [PDF] Chapter 2 - POISSON PROCESSES - MIT OpenCourseWare
    We now use the memoryless property of exponential rv's to find the distribution of the first arrival in a Poisson process after an arbitrary given time t > 0.
  66. [66]
    [PDF] 6.262 Lecture 4: Poisson (the perfect arrival process)
    Feb 14, 2011 · Since the interarrival interval for a Poisson process is exponential, i.e., Pr{X>x} = exp(−λx) for x > 0. exp(−λ(t + x)) = exp(−λt) exp(−λx).
  67. [67]
    [PDF] Lecture 2: Large deviations technique - MIT OpenCourseWare
    Sep 9, 2013 · Let us consider some examples of computing the moment generating functions. • Exponential distribution. Consider an exponentially distributed ...
  68. [68]
    [PDF] Survival Analysis
    Example: Exponential. • f(t) = λe-λt. • S(t) = e-λt. • h(t) = λ. S(t) h(t). 1 λ t t. Page 9. Weibull distribution. • Generalization of the exponential. • For λ, ...
  69. [69]
    [PDF] Markov Processes and Queues - MIT OpenCourseWare
    The exponential density function is therefore µe−µt for t ≥ 0 and 0 for t < 0. The time of the transition from 1 to 0 is said to be exponentially distributed.
  70. [70]
    Mathematical Treasure: Napier's Logarithms in English
    John Napier originally published his theory of logarithms in 1614 under the title, Mirifici Logarithmorum Canonis. The work was written in the scholarly ...
  71. [71]
    Charles Hermite (1822 - 1901) - Biography - MacTutor
    In 1873 Hermite published the first proof that e e e is a transcendental number. This is another result for which he is rightly famous. Using method's ...
  72. [72]
    Ferdinand von Lindemann (1852 - 1939) - Biography - MacTutor
    He is famed for his proof that π is transcendental, that is, π is not the root of any algebraic equation with rational coefficients. The problem of squaring the ...
  73. [73]
    Discussion on Errors - Python Programming And Numerical Methods
    We can estimate the error using the Taylor Remainder Estimation Theorem, which states: If the function f(x) has n+1 derivatives for all x in an interval I ...
  74. [74]
    Table-driven implementation of the Expm1 function in IEEE floating ...
    Algorithms and implementation details for the function ex - 1 in both single and double precision of IEEE 754 arithmetic are presented here.
  75. [75]
    A note on computing approximations to the exponential function
    These methods lead to approximations which are much superior to the commonly used convergents of the Gaussian continued fraction for the exponential.
  76. [76]
    [PDF] On the Implementation of Fixed-point Exponential Function ... - arXiv
    Look up tables and polynomial approximation have been used for implementing exponential and many non-linear functions [2,3]. Either of these two methods used in ...
  77. [77]
    FPGA implementation of exponential function using cordic IP core ...
    This paper proposes a method and a hardware architecture to calculate the exponential value of an input outside the convergence range of CORDIC IP core in ...