“…Therefore, without loss of generality, we can define a new random variable , and this has exactly the same statistics as , namely , with each component being a random complex Gaussian as before. Thus, the explicit integration algorithm that we employ reads (20) In the simulation of some toy model problems (not shown), this algorithm produced results for the distribution of intensity that were the same as integrating the complex equation of motion, whereas integrating the equivalent of (15), assuming a single noise source on the intensity, gave incorrect answers.…”