Abstract Firms often use a facility to produce both Make-to-Stock (MTS) and Make-to-Order (MTO) items. This study provides insights into conditions favoring the adoption of such a mixed MTS-MTO production strategy. Past studies ignore the influence of a one-to-many relation between semi-finished goods and products, and a guaranteed lead-time for MTO products, with uncertain product demand. The problem is modeled as a two-stage stochastic 0–1 integer, non-linear programming problem. In Stage 1, before demand realization, operations determine the product-wise mode of production and inventory level of semi-finished goods/MTS products, and marketing determine the guaranteed lead-time for MTO products. In Stage 2, product demand is realized, and it decreases linearly with lead-time and has a stochastic error component. The demand for MTS products is fulfilled from available inventory, and semi-finished goods are allocated to meet the demand of MTO products and build a stock of MTS products. The objective is to maximize the expected profit per-unit time. Sensitivity analysis of key problem parameters demonstrates that: (1) the integrated-firm can provide smaller lead-times under demand variability, (2) high holding cost of semi-finished goods, high lead-time sensitivity of demand, and high demand variability & low production rate promote mixed strategy over MTO strategy, and (3) high demand variability promotes substitution among semi-finished goods. A Hybrid Greedy and Non-linear Branch & Bound method (featuring problem-specific cuts and search & branching strategy) is proposed to solve real-life problems. This method provides a better solution and a smaller relative gap than commercial solver Simple Branch and Bound.

中文翻译：

---

通过混合按库存生产/按订单生产的生产策略管理产品品种，并在随机需求下保证交货时间

摘要 公司经常使用设施来生产按库存生产 (MTS) 和按订单生产 (MTO) 的产品。这项研究提供了有利于采用这种混合 MTS-MTO 生产策略的条件的见解。过去的研究忽略了半成品和产品之间一对多关系的影响，以及在产品需求不确定的情况下对 MTO 产品的保证提前期。该问题被建模为两阶段随机 0-1 整数非线性规划问题。在第一阶段，在需求实现之前，运营决定了半成品/MTS 产品的产品生产模式和库存水平，营销决定了 MTO 产品的保证提前期。在第 2 阶段，产品需求实现，它随着提前期线性减少并且具有随机误差分量。对 MTS 产品的需求从可用库存中得到满足，并分配半成品以满足 MTO 产品的需求并建立 MTS 产品的库存。目标是使每单位时间的预期利润最大化。关键问题参数的敏感性分析表明：（1）整合型企业可以在需求可变性下提供更短的提前期，（2）半成品持有成本高，需求提前期敏感性高，需求可变性高& 低生产率促进混合战略而不是 MTO 战略，(3) 高需求可变性促进半成品之间的替代。提出了一种混合贪婪和非线性分支定界方法（具有特定于问题的切割和搜索和分支策略）来解决现实生活中的问题。