计算机科学
排队论
排队
付款
激励
服务(商务)
约束(计算机辅助设计)
工作量
运筹学
动态定价
预算约束
数学优化
构造(python库)
一致性(知识库)
激励相容性
微观经济学
服务提供商
校长(计算机安全)
功能(生物学)
服务体系
分层排队网络
放弃(法律)
支付系统
队列管理系统
动态规划
作者
Tan Yu,Tiaojun Xiao,Xiang Zhong,Xiaohan Zhu
标识
DOI:10.1177/10591478261510178
摘要
Service systems in which demand, congestion, and worker incentives evolve jointly are ubiquitous in modern operations. We study a dynamic principal–agent problem in a queueing system with customer abandonment, where the principal controls arrivals through pricing and the agent privately chooses a costly service rate. Modeling workload as a controlled reflected diffusion, we characterize the centralized and decentralized Hamilton–Jacobi–Bellman equations and construct a representative two-part contract consisting of an access charge and a completion payment. Under unrestricted transfers, the contract coordinates pricing and service to attain the first-best joint payoff. Whenever participation is feasible, it makes the agent's participation constraint bind and allocates the remaining surplus to the principal. Implementation is robust to misspecification of the service-rate function used by the principal. We also establish conditions under which the representative payment schedules are nonmonotone in workload. Extensions consider multiple agents, general patience distributions, and information disclosure in a reduced Bayesian-persuasion benchmark. Numerical examples illustrate how the effects of customer impatience on optimal pricing and service can vary with congestion.
科研通智能强力驱动
Strongly Powered by AbleSci AI