Dynamic Matching in a Sharing Economy

本研究旨在研究共享经济中的动态匹配，以最大限度地提高社会总福利，通过（1）探索最佳政策的结构和（2）设计有效的可计算的启发式政策。 具体而言，我建议研究在共享经济中，在双方都存在跨期不确定性的情况下，如何使供给与需求相匹配。共享经济通常涉及的是一个双边市场，由供需方和中介做市商组成。有不同类型的供应和需求。匹配不同类型的供应和需求会产生不同的成本。中介组织/平台用于协调不同类型的供应和需求。例如，美国的器官共享联合网络（UNOS）将捐赠的器官分配给需要移植的患者。鉴于供应价格无关紧要（例如，在器官移植中）或外部给定的情况下，操作层面的匹配可以为平台提供跨不同需求类型跨时间分配众包供应的有效方式。共享平台的从业者需要量化的可实施工具来有效匹配供给和需求，考虑到随着时间的推移，会有新的供给和需求到达市场，不匹配的供给和需求可能会放弃等待。 在离散时间动态规划公式下，我计划（1）研究最优匹配策略的结构。然后，我计划（2）基于近似方法导出启发式策略，平台可以实现接近最优的社会福利最大化。对于某些启发式策略，我计划（3）导出它们的最坏情况性能界限。我建议将基本公式扩展到4个不同的方向，以解释（1）供应和需求的连续时间到达，（2）内生定价决策，（3）共享平台之间的竞争和（4）网络上的动态匹配。 我所提出的研究有望产生影响。首先，设计最优的跨期匹配策略将使平台实现更高的效率和社会福利。我计划在目前运营的共享平台上寻求可能的测试和实施衍生政策。第二，拟议的总框架建立在库存管理（侧重于供应方）和收入管理（侧重于需求方）这两个业务管理/研究的基础之上。鉴于我在顶级运营期刊上的出版记录，我有信心从这个研究计划（关于供需双方的动态匹配）中产生几篇高质量的论文，预计将吸引运营研究和管理科学界的许多后续工作。

This research aims to study dynamic matching in a sharing economy to maximize the total social welfare, by (1) exploring structures of optimal policies and (2) designing efficiently computable heuristic policies. To be specific, I propose to study the question of how to match supply with demand in a sharing economy under inter-temporal uncertainty on both sides. What is typically involved in a sharing economy is a two-sided market consisting of the supply and demand side and an intermediary market maker. There are different types of supply and demand. Matching different types of supply and demand incurs different costs. The intermediary organization/platform serves to coordinate supply and demand of different types. For example, United Network for Organ Sharing (UNOS) in the United States allocates donated organs to patients in need of transplantation. Given prices of supply are irrelevant (e.g., in organ transplantation) or exogenously given, matching at the operational level can provide an efficient way for the platform to allocate crowdsourced supply across different demand types inter-temporally. Practitioners of a sharing platform need quantitatively implementable tools to efficiently match supply and demand, taking into account that over time, there will be new arrivals of supply and demand to the market, and unmatched supply and demand may abandon the wait. Under a discrete-time dynamic programming formulation, I plan to (1) study the structures of the optimal matching policies. Then I plan to (2) derive heuristic policies, based on approximation methods, which the platform can implement to achieve near-optimal social welfare maximization. For certain heuristic policies, I plan to (3) derive their worst-case performance bounds. I suggest to extend the base formulation to 4 different directions, to account for (1) continuous-time arrivals of supply and demand, (2) endogenized pricing decisions, (3) competition between sharing platforms and (4) dynamic matching over a network. My proposed research is expected to make impacts. First, designing optimal inter-temporal matching policies would enable platforms to achieve higher efficiency and social welfare. I plan to seek possible test and implementation of the derived policies in currently operating sharing platforms. Second, the proposed general framework builds on inventory management (that focuses on the supply side) and revenue management (that focuses on the demand side), the two foundations of operations management/research. Given my publication record in the top operations journals, I am confident of having a couple of high-quality papers coming out of this research proposal (on dynamic matching of both the supply and demand side), which are expected to attract many follow-up works from the community of operations research and management science.