Suppose that π is a probability measure on the probability space \((S, \mathcal {A})\) , h is a measurable function from \(S\rightarrow \mathrm {R}\) , and one is interested in the calculation of the expectation \(\displaystyle \bar {h}=\int hd\pi \) assuming that the integral exists. In many problems, especially when the sample space S is multivariate or when the normalizing constant of π is not easily calculable, finding the value of this integral is not feasible either by numerial methods of integration (such as the method of quadrature) or by classical Monte Carlo methods (such as the method of rejection sampling).