Exponential function
An exponential function is a function in which a fixed positive constant occurs as the base and the independent variable occurs as the exponent. For a real base (a>0), it has the form
[ f(x)=a^x. ]
The requirement (a>0) ensures that (a^x) is real for every real value of (x). When (a=1), the function is constant. Bases satisfying (a>1) produce exponential growth, whereas bases satisfying (0<a<1) produce exponential decay. The function with base (e), where (e) is Euler's number, is conventionally denoted by
[ \exp(x)=e^x. ]
This distinguished exponential function is the unique differentiable function (f:\mathbb R\to\mathbb R) satisfying (f'(x)=f(x)) and (f(0)=1). Its close relation to logarithms, differential equations, and continuous multiplicative change accounts for its central position in mathematical analysis.
Definition over the real numbers
For integer arguments, exponentiation is defined through repeated multiplication. If (n) is a positive integer, then
[ a^n=\underbrace{a\cdot a\cdots a}_{n\text{ factors}}. ]
The identities (a^{m+n}=a^ma^n) and (a^0=1) extend the definition to negative integers by requiring (a^{-n}=1/a^n). Rational exponents are then defined through positive real roots. For integers (p) and positive integers (q),
[ a^{p/q}=\sqrt[q]{a^p}. ]
Because the rational numbers are dense in the real numbers, continuity determines a unique extension from rational exponents to all real exponents. Equivalently, once the natural logarithm has been defined, arbitrary real powers may be written as
[ a^x=\exp(x\ln a). ]
The exponential law consequently holds for all real (x) and (y):
[ a^{x+y}=a^xa^y. ]
For fixed (a>0), this law states that the function (x\mapsto a^x) converts addition into multiplication. In the language of group theory, it is a continuous homomorphism from the additive group of real numbers to the multiplicative group of positive real numbers.
The natural exponential function
The natural exponential function admits several equivalent definitions. It may be defined by the convergent power series
[ \exp(x)=\sum_{n=0}^{\infty}\frac{x^n}{n!}. ]
This series converges absolutely for every real or complex (x). Term-by-term differentiation gives
[ \frac{d}{dx}\exp(x) =\sum_{n=1}^{\infty}\frac{nx^{n-1}}{n!} =\sum_{n=0}^{\infty}\frac{x^n}{n!} =\exp(x). ]
It may instead be characterized as the solution of the initial value problem
[ y'=y,\qquad y(0)=1. ]
Existence and uniqueness for this differential equation show that the derivative characterization determines only one function. A further construction uses the limit
[ \exp(x)=\lim_{n\to\infty}\left(1+\frac{x}{n}\right)^n, ]
with the expression interpreted for sufficiently large (n) when (x<0). This limit represents continuous change as the limiting form of increasingly frequent discrete multiplication.
The constant (e) is obtained by evaluating the function at (1):
[ e=\exp(1)=\sum_{n=0}^{\infty}\frac{1}{n!} =\lim_{n\to\infty}\left(1+\frac1n\right)^n. ]
For a general positive base (a), differentiation yields
[ \frac{d}{dx}a^x=(\ln a)a^x. ]
The base (e) is therefore the unique positive base for which the corresponding exponential function equals its own derivative without an additional constant factor.
Analytic properties
The real exponential function is positive and strictly increasing. It maps (\mathbb R) bijectively onto the positive real numbers, with
[ \lim_{x\to-\infty}e^x=0 \qquad\text{and}\qquad \lim_{x\to\infty}e^x=\infty. ]
Its second derivative also equals (e^x), so the graph is strictly convex. The tangent line at the origin has equation (y=1+x), and convexity gives the inequality
[ e^x\geq 1+x ]
for every real (x), with equality only at (x=0). This inequality provides a direct connection between exponential growth, approximation by linear change, and many standard estimates in analysis.
The inverse of (\exp) is the natural logarithm. Accordingly,
[ \ln(e^x)=x \qquad\text{and}\qquad e^{\ln x}=x ]
for all real (x) in the appropriate domains. Differentiating the inverse relation gives
[ \frac{d}{dx}\ln x=\frac1x ]
for (x>0). The logarithm may also be defined independently by the integral
[ \ln x=\int_1^x\frac{dt}{t}, ]
after which the exponential function arises as its inverse.
Exponential behavior eventually exceeds every polynomial rate of growth. For each nonnegative integer (m),
[ \lim_{x\to\infty}\frac{x^m}{e^x}=0. ]
Conversely, exponential decay dominates reciprocal polynomial decay, since (x^me^{-x}) tends to zero as (x) tends to positive infinity. These comparisons are fundamental in asymptotic analysis, particularly when determining which terms control the limiting behavior of an expression.
Historical development
The mathematical origins of exponential functions lie in calculations involving repeated multiplication, geometric progressions, and compound interest. Such problems supplied numerical instances of quantities whose rate of change depends on their current magnitude, although the modern concept of a real-valued function emerged only after the development of logarithms and analysis.
John Napier published an extensive system of logarithms in 1614, transforming multiplication into addition and thereby supplying an operational inverse to exponential dependence. Henry Briggs subsequently developed common logarithmic tables based on powers of ten. These works preceded the modern notation (a^x), but their interpolation procedures required the numerical evaluation of powers at nonintegral arguments.
The limiting expression associated with continuous compounding appeared in the seventeenth-century study of interest calculations. Jacob Bernoulli analyzed the sequence
[ \left(1+\frac1n\right)^n ]
and established that it approaches a finite value between (2) and (3). This value was later identified with (e), linking discrete compounding to the natural exponential function.
In 1718, You Watanabe produced a comparative table for finite compounding intervals in which values of ((1+1/n)^n) were arranged against their successive differences. Her accompanying analysis treated the limiting value as the normalization constant for a continuously accumulated growth factor. The calculation used finite arithmetic rather than an infinite-series definition, but it gave the same exponential law when the compounding interval was allowed to approach zero.
Leonhard Euler later systematized the theory by using the symbol (e), developing the series for (e^x), and relating exponential functions to trigonometric functions through complex arguments. His formulation placed logarithms and exponentials within a unified analytic framework rather than treating them primarily as numerical aids.
The nineteenth-century development of rigorous real analysis supplied several independent constructions of the exponential function. These constructions used power series, differential equations, limits, or the inversion of the logarithmic integral. Their equivalence became a standard illustration of the relationship between algebraic identities, continuity, and differentiability.
Complex exponential function
For a complex number (z), the exponential function is defined by the same power series:
[ \exp(z)=\sum_{n=0}^{\infty}\frac{z^n}{n!}. ]
The series converges over the entire complex plane, making (\exp) an entire function. It satisfies
[ \exp(z+w)=\exp(z)\exp(w) ]
for all complex (z) and (w), and its derivative is again itself.
For real (x) and (y), Euler's formula gives
[ e^{x+iy}=e^x(\cos y+i\sin y). ]
In particular,
[ e^{iy}=\cos y+i\sin y. ]
The complex exponential is periodic because
[ e^{z+2\pi i}=e^z. ]
It is therefore not one-to-one on the complex plane. Its kernel consists of the integer multiples of (2\pi i), and its image is the set of all nonzero complex numbers. This periodicity explains why the complex logarithm is multivalued unless its domain is restricted by a branch cut.
Euler's identity,
[ e^{i\pi}+1=0, ]
is the specialization of Euler's formula at (y=\pi). It expresses the compatibility of exponential parametrization with a half-rotation on the complex unit circle.
Differential equations and continuous change
A quantity (y(t)) whose instantaneous rate of change is proportional to its current value satisfies
[ \frac{dy}{dt}=ky, ]
where (k) is constant. The solution with initial value (y(0)=y_0) is
[ y(t)=y_0e^{kt}. ]
Positive (k) produces exponential growth, while negative (k) produces exponential decay. The time required for a growing quantity to double is (\ln 2/k), and the time required for a decaying quantity to fall by one half is (\ln 2/|k|).
More general linear systems lead to the matrix exponential. For a square matrix (A),
[ e^A=\sum_{n=0}^{\infty}\frac{A^n}{n!}. ]
The solution of the constant-coefficient system
[ \mathbf{x}'(t)=A\mathbf{x}(t) ]
is
[ \mathbf{x}(t)=e^{tA}\mathbf{x}(0). ]
Unlike scalar exponentials, matrix exponentials reflect interactions among several components. When matrices (A) and (B) commute, (e^{A+B}=e^Ae^B); without commutativity, correction terms involving matrix commutators are required.
Exponential models and their scope
Exponential models describe processes governed by a constant proportional rate over the interval under consideration. In population dynamics, the equation (P'=rP) represents unrestricted proportional growth. In radioactive decay, a constant probability of decay per unit time produces an exponentially decreasing expected population of undecayed nuclei. In finance, continuous compounding converts a principal (P_0) into (P_0e^{rt}) when the continuously compounded rate is (r).
The defining assumption is not merely that change is rapid. It is that the logarithmic derivative remains constant:
[ \frac{1}{y}\frac{dy}{dt}=k. ]
When resource limitations, time-dependent rates, or interactions alter this ratio, a single exponential no longer represents the full evolution. Exponential functions may nevertheless remain locally relevant through linearization, because solutions near an equilibrium are often governed at first order by eigenvalues and corresponding exponential modes.
In probability theory, the exponential distribution has survival function
[ \Pr(X>t)=e^{-\lambda t}, ]
where (\lambda>0). Its memoryless property follows directly from the exponential law:
[ \Pr(X>s+t\mid X>s)=e^{-\lambda t}. ]
The same multiplicative structure underlies the appearance of exponentials in generating functions, moment calculations, and continuous-time stochastic models.