首先利用控制工程领域的方法描述了单阶供应链中基于AR(1)需求和最小均方误差预测的DE-APVIOBPCS模型。通过应用终值定理可知,不论提前期的估计准确与否,系统的终值偏差(即库存偏移)都将始终存在。为了消除此系统的固有偏差并保持合理的系统动态性,作者提出了一种基于DE-APVIOBPCS模型的零库存偏移变体。由方差放大分析可知,在新的变体模型中采用保守的提前期估计值总是比较有利。分析了零库存偏移变体模型的稳定性条件,并在提前期错误估计的假设前提下通过仿真分析举例说明了新模型的一些优点。
The DE-APVIOBPCS model with MMSE forecast for AR( 1 ) demand in a single-echelon supply chain has first been described in control engineering perspective. By applying the Final Value Theorem,a final value offset (i. e. inventory drift) can be measured and does exist even though the actual lead-time is known. Thus to eliminate the inherent offset and keep the system variances acceptable,a new policy with zero inventory drift based on DE-APVIOBPCS model is presented . The analysis of the variance amplification suggests the lead-times conservatively in new policy should be always estimated. The general stability conditions for zero inventory drift variant are evaluated in succession and some valuable attributes of new policy are illustrated via simulation under the assumption that misidentification of lead-time is inevitable.