This study addresses a multi-period product-line-mix problem considering product interdependencies and uncertainties regarding price, demand, production cost, and the cannibalization effect of new products. The problem is modeled as a two-stage stochastic program. In the first-stage, decisions for the release times of new product-lines and capacity expansion are made without waiting for the realization of random events. Sales volumes are determined in the second-stage after more information about uncertainties is revealed. The solution approach employs a sample average approximation technique based on the Monte Carlo bounding, and multi-cut version of the L-shaped method to solve approximate problems efficiently. The model and solution approach are tested on different cases considering the value-of-stochastic-solution (VSS) and the expected-value-of-perfect-information (EVPI) as performance measures, under the assumption that the decision-maker is risk-neutral. Data collected through experimental studies are analyzed using ANOVA and Random Forest method to determine the impact of both deterministic and uncertain parameters on the performance of the stochastic approach, and to generate some rule-based inferences about the behavior of the proposed model. The results demonstrate the problem environments in which the stochastic model can generate the highest benefit to the business. It is mainly found the model can provide more expected profit than the mean-value solution in problems having uncertainties, particularly when selling prices and production costs are uncertain.

本研究解决了一个多期产品线组合问题，考虑到产品在价格、需求、生产成本和新产品的蚕食效应方面的相互依赖性和不确定性。该问题被建模为一个两阶段的随机程序。第一阶段，无需等待随机事件的实现，就可以决定新产品线的发布时间和产能扩张。在披露更多不确定性信息后，在第二阶段确定销售量。解决方法采用基于蒙特卡罗边界的样本平均近似技术，以及L 的多切割版本形方法有效地解决近似问题。在假设决策者是风险的情况下，将随机解决方案的价值 (VSS) 和完美信息的预期价值 (EVPI) 作为绩效衡量标准，对模型和解决方案方法进行了测试-中性的。使用方差分析和随机森林方法分析通过实验研究收集的数据，以确定确定性和不确定性参数对随机方法性能的影响，并生成一些关于所提出模型行为的基于规则的推断。结果证明了随机模型可以为业务带来最大收益的问题环境。