Stochastic process

A stochastic process, often called a random process, is a mathematical object used to model systems that change in unpredictable ways. Instead of following a single, fixed path, these processes represent phenomena that vary due to chance. In probability theory, a stochastic process is defined as a family of random variables. These variables are organized within a probability space. The index of this family is known as the index set, and it often represents the passage of time. By using these models, scientists can study complex movements like the growth of bacterial populations or the flickering of electrical currents caused by thermal noise.

To understand how these processes function, we must look at their internal structure. Each random variable in the collection is uniquely linked to an element in the index set. The values that these variables take belong to a mathematical area called the state space. This space can be simple, such as the set of integers, or more complex, like $n$-dimensional Euclidean space. When we measure the change in a process between two specific points in the index set, we call that change an increment. Because the process is driven by randomness, it can produce many different possible outcomes. A single, specific outcome from the process is called a realization or a sample function.

Mathematicians classify stochastic processes using several different criteria. One common method is to look at the cardinality, or size, of the index set and the state space. If the index set consists of a finite or countable number of elements, such as the natural numbers, the process is in discrete time. These are often called random sequences. If the index set is an interval of the real line, the process is in continuous time. Discrete-time processes are generally easier to study than continuous-time processes. This is because continuous-time processes require advanced mathematical tools to handle an uncountable index set.

The state space also determines the name of the process. If the state space is made of integers, it is a discrete or integer-valued process. If the state space is the real line, it is a real-valued process with a continuous state space. When the state space involves higher dimensions, it is called a vector process. If the index set is a higher-dimensional Euclidean space rather than a line, the collection of variables is called a random field. These classifications allow researchers to apply the correct mathematical frameworks to specific real-world problems.

The history of these ideas involves centuries of linguistic and mathematical evolution. The word "stochastic" comes from a Greek word meaning "to aim at a mark" or "to guess." In 1713, Jakob Bernoulli used the phrase "Ars Conjectandi sive Stochastice" in his work on probability. The specific English term "stochastic process" appeared in a 1934 paper by Joseph Doob.

Doob cited a German paper by Aleksandr Khinchin, who had used the term "stochastischer Prozeß." Other mathematicians, such as Andrei Kolmogorov, had used similar terms as early as 1931. These developments helped formalize the study of randomness.

Several classic processes serve as pillars for the field. The Wiener process, also known as Brownian motion, is one of the most central models. It is a continuous-time process where the increments are normally distributed. Norbert Wiener proved its mathematical existence.

Historically, Louis Bachelier used the Wiener process to model price changes on the Paris Bourse. Another essential model is the Poisson process, which A. K. Erlang used to model the number of phone calls occurring in a specific period. There are also random walks, which can be discrete or continuous, representing movements that change based on random steps.

Stochastic processes are vital across many scientific disciplines. In biology, they model population changes, while in physics, they describe the movement of gas molecules. Finance relies heavily on these models to understand the seemingly random fluctuations of markets. They are also used in neuroscience, ecology, and chemistry. Even modern technology uses them in image processing, signal processing, and telecommunications. By using tools from calculus, linear algebra, and measure theory, mathematicians continue to expand our understanding of the random world.