发表机构
Chung Yuan Christian University（中原大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文研究人口冲击下稳定匹配的历史可达性，证明仅终端主体历史无法覆盖全部稳定匹配，而通过保持偏好的临时主体扩充历史可恢复完整稳定集合。
AI 中文摘要
序贯进入程序不必达到每一个稳定匹配。我们探究当双方主体可能反复进入和退出，且每次人口变化后通过提议链重新稳定化时，完全可达性是否得以恢复。答案是否定的：一个三乘三的婚姻市场存在一个稳定匹配，该匹配在仅限终端主体的任何此类历史中，从空活跃市场出发是不可达的。随后，我们允许在终端日期缺席的临时历史主体。对于每个目标稳定匹配，我们构造一个依赖于目标的、保持偏好的扩充，以及一条可达该匹配的容许历史，其中所有临时主体在终止前退出。该构造可推广到具有响应性医院偏好的多对一市场。因此，可达性取决于容许历史的类别。终端主体历史可能选择稳定集合的一个真子集，而保持偏好的扩充历史则恢复整个稳定集合。当此类扩充历史是容许的但过去的参与者不可观测时，仅凭终端原始数据并不能排除任何稳定匹配。
英文摘要
Sequential-entry procedures need not reach every stable matching. We ask whether full reachability is restored when agents on both sides may enter and exit repeatedly, with proposal-chain restabilization after each population change. It is not: a three-by-three marriage market has a stable matching that is unreachable from the empty active market along any such history restricted to terminal agents. We then allow temporary historical agents who are absent at the terminal date. For every target stable matching, we construct a target-dependent, preference-preserving augmentation and an admissible history that reaches it, with all temporary agents exiting before termination. The construction extends to many-to-one markets with responsive hospital preferences. Thus reachability depends on the admissible history class. Terminal-agent histories may select a proper subset of the stable set, whereas preference-preserving augmented histories recover the full stable set. When such augmented histories are admissible but past participants are unobserved, the terminal primitives alone do not exclude any stable matching.
