Additionrandom

Additionrandom is a term used in some discussions to describe a class of randomness-generating processes in which a random output is produced by adding together multiple random inputs. Although not a widely standardized term in mathematics, it is used to describe two related ideas: additive construction of random variables via sum, and algorithmic schemes that produce randomness by successive addition of random components.

Definition: Let X1, X2, ..., Xn be independent random variables with respective distributions F1, F2, ..., Fn. The random variable S_n = X1 + X2 + ... + Xn is called an additionrandom output in this informal sense. The distribution of S_n is the n-fold convolution of Fi. When weights w_i are used, S = sum w_i Xi defines a generalized additionrandom.

Variants: modulo additionrandom computes S mod m; this can create different discrete distributions depending on the Xi. When the Xi are Bernoulli or uniform, one obtains familiar distributions (binomial, discrete uniform) in the sum.

Properties: by the central limit theorem, under mild conditions, S_n tends toward a normal distribution as n grows. The exact distribution is given by convolution; adding more components smooths the distribution but preserves the general idea of accumulated randomness.

Applications: modeling accumulated random shocks, constructing random number generators via additive schemes, and randomized algorithms where several independent sources of randomness are combined.

Limitations: independence of inputs must be ensured, the resulting distribution should be analyzed, and biases in finite sums should be accounted for.

See also: sum of independent random variables, convolution, central limit theorem, random walk.