A mathematical programming-based solution method for the nonstationary inventory problem under correlated demand

This paper extends the single-item single-stocking location nonstationary stochastic inventory problem to relax assumption of independent demand. We present a mathematical programming-based solution method built upon an existing piecewise linear approximation strategy under receding horizon control framework. Our can be implemented by leveraging off-the-shelf mixed-integer programming solvers. It tackle demand various assumptions: multivariate normal distribution, collection time-series processes, and Martingale Model Forecast Evolution. compare against exact solutions obtained via dynamic demonstrate that our leads near-optimal plans.