Bound estimation of multistage asset allocation based on MCMC

Stochastic programming models in multistage asset allocation almost always require the calculation of some sort of expectation in the form of integral and its optimization. This paper is to provide a numerical solution of multidimensional integrals via Markov Chain Monte Carlo (MCMC) integration. The method proposed in this paper can alleviate the major drawback of multistage stochastic programming, i.e. the exponential growth of scenarios. The numerical simulation support our view.