Non-Markovian stochastic jump processes. I. Input field analysis.

Abstract

A non-Markovian model of correlated phase jumps is introduced for phase fluctuations of an electromagnetic field. This generalized jump model (GJM) treats phase jumps of arbitrary size, occurring at random times; but in contrast to previous work, the jumps are allowed to be fully correlated, partially correlated, or uncorrelated. The degree of correlation is defined by a single parameter derived from the theory. The familiar phase-diffusion model, telegraph-noise model, Burshtein model, and Brownian-motion-like model are all obtained from the GJM in the proper limits. The standard way of characterizing the spectrum of a laser has been the assignment of a single parameter---the linewidth. However, in experiments where the details of the fluctuations are important, or where exact line shapes are measured, this single-parameter characterization might be insufficient. This GJM describes most cases by a set of three stochastic parameters: the degree of correlation between the jumps, the characteristic jump size, and the mean time between jumps. In this paper expressions are derived for the correlation function and the spectrum of a stochastic field in terms of these three stochastic parameters. In addition to analytical work, detailed numerical simulations are presented for the various limiting cases of the model, and the agreement between theory and simulation is excellent. Since the stochastic parameters are not a priori known, a procedure is described for extracting the stochastic parameters from measurable quantities such as the field correlation function or spectrum. Since correlated fluctuations are very common in optics (any stabilization feedback procedure involves anticorrelation), the questions of relevance of the present model to problems of current interest in optical communication and nonlinear optics are also discussed.