Sequential Monte Carlo methods are a class of simulation-based algorithms which provide sample approximations of each of a sequence of distributions in turn, as well as their corresponding normalizing constants. They find widespread application in online inference for time series but are much more generally applicable.