Godfrey, G. and W.B. Powell, “An Adaptive, Distribution-Free Algorithm for the Newsvendor Problem with Censored Demands, with Application to Inventory and Distribution Problems,” Management Science, Vol. 47, No. 8, pp. 1101-1112, (2001). (c) Informs.
This paper proposes a novel approximation procedure for stochastic resource allocation problems. The technique works with a series of stochastic gradients which are then used to maintain a concave estimate of the value function. When applied to problems with known optimal solutions, the method provides results within a fraction of a percent of optimal. One byproduct is a fast algorithm for the newsvendor problem that does not require knowledge of the underlying demand distribution. The technique is also applied to separable distribution problems.
(Click here to download paper)
“Models and Algorithms for Distribution Problems with Uncertain Demands,” Transportation Science, Vol. 30, No. 1, pp. 43-59 (with R. K.-M. Cheung). (c) Informs
(Click here to download paper)