Exponential distribution
Generated by GPT-5-miniExpansion Funnel Raw 56 → Dedup 11 → NER 11 → Enqueued 7
| Exponential distribution | |
|---|---|
| Name | Exponential distribution |
| Type | Continuous |
| Support | [0, ∞) |
| Parameters | Rate (λ) or scale (β) |
| Mean | 1/λ |
| Variance | 1/λ² |
Exponential distribution The exponential distribution is a continuous probability distribution commonly used to model waiting times and lifetimes in queuing and reliability contexts. Its mathematical form and properties tie it to foundational work by Andrey Kolmogorov, Andrei Markov, Aleksandr Lyapunov, André-Louis Cholesky, and applications in studies by Agner Krarup Erlang, Harald Cramér, William Feller, Ronald Fisher, and Jerzy Neyman.
Definition and basic properties
The distribution is defined on the nonnegative real line with a single positive rate parameter, often denoted λ, and a scale parameter β = 1/λ; foundational treatments appear in texts associated with Kolmogorov and Feller. Its support [0, ∞) and monotone decreasing density relate to limit results by Paul Lévy, S. N. Bernstein, and connections to the Poisson process developed by Siméon Denis Poisson, A. K. Erlang, Agner Krarup Erlang, and later formalized by A. N. Kolmogorov. The distribution is closed under minima of independent exponentials, a property exploited in works by Harald Cramér, William Feller, and Andrey Markov when analyzing renewal processes such as those in Erlang's loss model and queueing models studied by John von Neumann and Claude Shannon.
Probability density and cumulative distribution
The probability density function f(x) = λ e^{-λ x} for x ≥ 0 and cumulative distribution function F(x) = 1 − e^{-λ x} are standard in treatments by Ronald Fisher, Jerzy Neyman, Harald Cramér, William Feller, and Andrey Kolmogorov. These expressions are used in proofs by Paul Lévy, Marcel Riesz, and Norbert Wiener when connecting exponential laws to stochastic processes such as the Poisson process introduced by Siméon Denis Poisson and extended in work by Andrey Kolmogorov and Andrei Markov. The hazard function h(x) = λ is constant, a feature emphasized in reliability analyses by W. Edwards Deming, Shewhart, and Walter A. Shewhart-inspired industrial statistics, and in survival studies associated with Kaplan–Meier methodology and Bradford Hill-style epidemiology.
Moments and characteristic functions
Moments of order n are given by n!/λ^n, a result discussed in classical texts by Harald Cramér, William Feller, Ronald Fisher, and Jerzy Neyman. The mean 1/λ and variance 1/λ² appear in derivations by Andrey Kolmogorov and Paul Lévy, while the moment-generating function M(t) = λ/(λ − t) for t < λ and characteristic function φ(t) = λ/(λ − it) are used in transform methods by Marcel Riesz, Norbert Wiener, and Salomon Bochner. Higher cumulants and factorial moments are treated in probabilistic analyses by Pafnuty Chebyshev and S. N. Bernstein and are instrumental in limit theorems influenced by Andrey Kolmogorov and Aleksandr Lyapunov.
Memoryless property and related distributions
The exponential distribution is the unique continuous distribution with the memoryless property, a characterization paralleling discrete results for the geometric distribution studied by Jakob Bernoulli, Abraham de Moivre, and formalized in later expositions by William Feller and Ronald Fisher. This property underpins connections to the Poisson process of Siméon Denis Poisson and renewal theory developed by Agner Krarup Erlang and Jesse Douglas, and it leads to related families such as the gamma distribution (sum of exponentials) explored by Karl Pearson and Ronald Fisher, and the Weibull distribution examined by Waloddi Weibull and applied by Frank Ramsey and Harald Cramér in reliability contexts.
Parameter estimation and inference
Common estimators for the rate λ include the maximum likelihood estimator and Bayesian posterior summaries; classical derivations appear in works by Ronald Fisher, Jerzy Neyman, Egon Pearson, and Harold Jeffreys. The likelihood function for independent exponential samples and the corresponding confidence intervals are treated in Neyman-Pearson theory and in Fisherian likelihood analyses, while conjugate prior analysis using the gamma family is standard in Bayesian expositions by Thomas Bayes, Pierre-Simon Laplace, Harold Jeffreys, and modern texts influenced by Dennis Lindley and Bradley Efron. Hypothesis tests and goodness-of-fit methods reference techniques popularized by Karl Pearson, William Gosset (Student), and Andrey Kolmogorov's tests.
Applications and examples
Exponential models are widely used in survival analysis in clinical trials associated with Austin Bradford Hill, in reliability engineering in work by W. Edwards Deming and Walter A. Shewhart, and in telecommunications building on models by Agner Krarup Erlang, Claude Shannon, John von Neumann, and Norbert Wiener. They model interarrival times in queueing systems in studies by Erlang and D. G. Kendall, lifetimes in actuarial science influenced by Edmund Halley and Benjamin Gompertz, and decay processes in physics linked to empirical laws discussed by Radioactive decay pioneers such as Ernest Rutherford and Marie Curie. Practical examples include modeling service times in call centers studied in operations research by Herbert Simon and George Dantzig, failure times in aerospace engineering projects associated with NASA analyses, and waiting times in network traffic examined by Vinton Cerf and Tim Berners-Lee.