The Nonstationary Stochastic Lead-Time Inventory Problem: Near-Myopic Bounds, Heuristics, and Testing
结合经典随机前置期库存模型与近期零滞后研究，为非平稳随机前置期问题提出近视界边界和启发式算法，测试表明近视启发式在中等变化问题中误差低于1.5%，而近近视启发式平均误差仅0.5%。
The purpose of the current paper is to combine the classical results of Kaplan (Kaplan, R. 1970. Dynamic inventory model with stochastic lead times. Management Sci. 16(2) 491–507.) and Ehrhardt (Ehrhardt, R. 1984. (s, S) Policies for a dynamic inventory model with stochastic lead times. Oper. Res. 32(1) 121–132.) for stochastic leadtime problems with recent work of Morton and Pentico (Morton, T., D. Pentico. 1995. The finite horizon nonstationary stochastic inventory problem near-myopic bounds, heuristics, testing. Management Sci. 41(2) 334–343.), which assumed zero lag, to obtain near-myopic bounds and heuristics for the nonstationary stochastic leadtime problem with arbitrary sequences of demand distributions, and to obtain planning horizon results. Four heuristics have been tested on a number of different demand scenarios over a number of random trials for four different leadtime distributions. The myopic (simplest) heuristic performs well only for moderately varying problems without heavy end of season salvaging, giving errors for this type of problem that are less than 1.5%. However, the average error for the myopic heuristic over all scenarios tested is 20.0%. The most accurate heuristic is the near-myopic heuristic which averages 0.5% form optimal across all leadtime distributions with a maximum error of 4.7%. The average error with increases in variance of the leadtime distribution.