Multi-dimensional Newsvendor (MDNV)

This is an instance from http://simopt.org/wiki/index.php?title=Multi-dimensional_Newsvendor_Problem. The example is adapted from the article Kim, S. (2006). Gradient-Based Simulation Optimization. Proceedings of the 2006 Winter Simulation Conference, 159-167. For this instance, we compare the performance of different stochastic algorithms: Stochastic Approximation(SA), Robust Stochastic Approximation(RSA) and Stochastic Decomposition(SD). To run the experiment of SA/RSA, please download the input file(model file for the subproblem) and the program files. Both algorithms run 10 replications, while all the parameters are setup in the corresponding config file. To run SD, please download the SMPS file and compile the SD code. The parameters are setup in the config.sd file.