Stochastic Programming Method for Volatile Resource Distribution Problem

Resource distribution is an important problem in many fields. It is particularly important when the supplied resource is volatile and the individual distinct demands are stochastic. In such cases, the uncertainty of demands must be put into consideration for both economic and resource efficiency. We first present a practical case of such problem with a fuel cell plant waste heat energy utilization issue, then generalize this problem into a stochastic programming problem, and develop a numerical framework that can determine the distribution profile for optimal efficiency. We compare experimental results with distribution profile computed by deterministic algorithm. The stochastic method yields greater expected revenue in more than 99% of cases with average improvement of 1% over the deterministic method. We also discuss the computation time of our method.