| 共生换代场景下持续创新型产品的多周期联合决策研究 |
| Joint decision on sequentially innovative products: A symbiotic replacement perspective |
| 摘要点击 37 全文点击 0 投稿时间：2024-09-21 修订日期：2026-01-06 |
| 查看/发表评论 下载PDF阅读器 |
| 中文关键词 以旧换新；共生换代；多周期联合决策；深度强化学习 |
| 英文关键词 Trade-in; Symbiotic replacement; Multi-period joint decision-making; Deep reinforcement learning |
| 基金项目 国家社会科学基金(22BGL289) |
| 投稿方向 产品运营决策 |
| 作者 | 单位 | 邮编 | | 郑江波 | 暨南大学管理学院 | 510632 | | 叶增健 | 中国外运华南有限公司 | | | 陈可嘉* | 暨南大学管理学院 | 510520 |
|
| 中文摘要 |
| 面向现实中以旧换新最常见的两代产品共生换代场景，研究了多周期下新产品与现有产品的定价及上架数量，以及旧产品回收价格的联合决策问题.基于马尔可夫决策过程理论，构建了一个多周期的联合决策模型，采用深度强化学习算法，通过数据仿真实验得出了全局近似最优策略，并采用可解释性框架深入分析了影响厂商最优决策收益的关键因素.结果表明: 厂商基于有限的状态与决策选择，能够有效获得多周期下近似最优的联合策略，从而实现整体收益的近似最大化；其次，将持有不同产品的消费者群体的细分情况，以及每期销售剩余的新产品数量作为决策依据，可以有效提升厂商的整体性收益；再次，现有产品的动态定价与新产品的上架数量对于厂商的整体收益的影响最为重要；最后，旧产品回收价格的提升有助于以旧换新，而且有利于提升消费者的忠诚度. |
| 英文摘要 |
| This study investigates the multi-period joint decision-making problem involving pricing, quantity of new and existing products, and trade-in value for old products during a two-generation sequentially innovative product transition. A multi-period joint decision model is developed based on Markov Decision Process theory and solved using deep reinforcement learning. Simulation experiments derive a globally near-optimal strategy. Furthermore, an explainable framework identifies key factors influencing profits. Results indicate that manufacturers can achieve near-optimal overall profits with limited states and decisions. Segmenting consumer groups by product ownership and considering leftover new product inventory significantly enhance profits. Dynamic pricing of the existing product and the quantity of new products are the most critical factors. Increasing the trade-in value promotes upgrades and strengthens customer loyalty. |
| 关闭 |
