Ergodic process | EPFL Graph Search

In physics, statistics, econometrics and signal processing, a stochastic process is said to be in an ergodic regime if an observable's ensemble average equals the time average. In this regime, any collection of random samples from a process must represent the average statistical properties of the entire regime. Conversely, a process that is not in ergodic regime is said to be in non-ergodic regime. One can discuss the ergodicity of various statistics of a stochastic process. For example, a wide-sense stationary process has constant mean and autocovariance that depends only on the lag and not on time . The properties and are ensemble averages (calculated over all possible sample functions ), not time averages. The process is said to be mean-ergodic or mean-square ergodic in the first moment if the time average estimate converges in squared mean to the ensemble average as . Likewise, the process is said to be autocovariance-ergodic or d moment if the time average estimate converges in squared mean to the ensemble average , as . A process which is ergodic in the mean and autocovariance is sometimes called ergodic in the wide sense. The notion of ergodicity also applies to discrete-time random processes for integer . A discrete-time random process is ergodic in mean if converges in squared mean to the ensemble average , as . Ergodicity means the ensemble average equals the time average. Following are examples to illustrate this principle. Each operator in a call centre spends time alternately speaking and listening on the telephone, as well as taking breaks between calls. Each break and each call are of different length, as are the durations of each 'burst' of speaking and listening, and indeed so is the rapidity of speech at any given moment, which could each be modelled as a random process. Take N call centre operators (N should be a very large integer) and plot the number of words spoken per minute for each operator over a long period (several shifts). For each operator you will have a series of points, which could be joined with lines to create a 'waveform'.

On The Convergence Of Stochastic Primal-Dual Hybrid Gradient

Volkan Cevher, Ahmet Alacaoglu

In this paper, we analyze the recently proposed stochastic primal-dual hybrid gradient (SPDHG) algorithm and provide new theoretical results. In particular, we prove almost sure convergence of the iterates to a solution with convexity and linear convergenc ...

SIAM PUBLICATIONS2022