Abstract

Shortening the lead time for Product Development (PD) provides enterprises with a competitive advantage. Given the iterative nature of PD projects, two aspects are regularly considered to shorten the PD lead time, that is, conducting faster or fewer iterations. However, executing faster iterations usually causes more iterations and vice versa. Therefore, suitable coordination between faster and fewer iterations is necessary to minimize the PD lead time. We investigate this coordination from a strategic perspective, whereby a PD project is considered as a sequence of stages and characterized by the design rates and rework probabilities of those stages. We model the coordination as a decision to choose the appropriate design rates for each stage, wherein the rework probabilities are negatively related to the design rates. An absorbing Markov process is applied to calculate the expected lead time of a PD project. Further, we formulate a geometric programming model to determine the optimal design rates of the stages with respect to the minimal expected lead time. Several insights are extracted from the model to provide general guidance on the coordination, including the effect of the acceptance check rate of the project, rework risk of the stages on the optimal design rates, and decomposability of the coordination. Inspired by these insights, an efficient heuristic algorithm is designed. The algorithm performs well in numerical experiments, which in turn validates the insights. Additionally, a field case proves the effectiveness of our model. Compared with the current policy, 12.25% of the PD lead time is saved through appropriate coordination between faster and fewer iterations.

摘要

缩短产品开发 (PD) 的前置时间可为企业提供竞争优势。鉴于 PD 项目的迭代性质，通常会考虑两个方面来缩短 PD 提前期，即进行更快或更少的迭代。然而，执行更快的迭代通常会导致更多的迭代，反之亦然. 因此，需要在更快和更少的迭代之间进行适当的协调，以最大限度地减少 PD 提前期。我们从战略角度研究这种协调，将 PD 项目视为一系列阶段，并以这些阶段的设计率和返工概率为特征。我们将协调建模为为每个阶段选择适当设计率的决定，其中返工概率与设计率呈负相关。应用吸收马尔可夫过程来计算 PD 项目的预期提前期。此外，我们制定了一个几何规划模型，以确定与最小预期提前期相关的阶段的最佳设计率。从模型中提取了一些见解，以提供有关协调的一般指导，包括项目验收率的影响、阶段返工风险对最优设计率的影响，以及协调的可分解性。受这些见解的启发，设计了一种有效的启发式算法。该算法在数值实验中表现良好，进而验证了见解。此外，现场案例证明了我们模型的有效性。与当前策略相比，通过更快和更少迭代之间的适当协调，可节省 12.25% 的 PD 提前期。一个现场案例证明了我们模型的有效性。与当前策略相比，通过更快和更少迭代之间的适当协调，可节省 12.25% 的 PD 提前期。一个现场案例证明了我们模型的有效性。与当前策略相比，通过更快和更少迭代之间的适当协调，可节省 12.25% 的 PD 提前期。