Eppen構建了單一貨品、單期、多個銷售站之多站報童決策模式，在最佳存貨水準下，比較最佳分散及最佳集中存貨政策總期望成本之大小。然而，最佳分散及最佳集中存貨水準之大小，又如何呢？是值得探討的問題。因此，本文假設：（1）銷售站數n個；（2）各銷售站之隨機需求量具有相同的常態分配N（μ，σ2）；（3）各銷售站均具有相同的線性持有成本函數h及線性缺貨成本函數p。針對參數n，p，h，μ與σ，應用MathCAD數值軟體，對Eppen所建構的分散及集中存貨模式之最佳分散存貨水準S* 及最佳集中存貨水準S0*作一比較，經4,000次模擬，得到主要結果是：最佳分散存貨水準S*大於最佳集中存貨水準S0*成立之充要條件為，單伅缺貨成本p大於單位持有成本h。這個重要性質，可以做為決策者設計存貨政策之參考。

　　　　　Eppen constructed a decision model of Mulit-location newsboy problem to investigate a single-product, single-period, multi-location inventory policy selection problem. To minimize the expected total cost, inventory levels in different policies will be considered. The assumptions are: (1) the munber of the locations is n: (2) demand of each location has an independent and identical normal distribution N (μ，σ2) : (3) the linear holding and penalty cost functions h、p at each location are assumed to be identical. Under the conditions that n, p, h, μand σ are fixed, the MathCAD numerical software was applied to compare the optimizing inventory levels of Eppen's models. Through 4,000trials of experimental simulation, the result obtained is: Let S* and S0* be the optimal inventory levels of decentralized system and centralized system, respectively. Then S* > S0* if and only if p>h. Which can be valuable reference for inventory policy makers.