共享经济预约模式下资源分配的拍卖机制设计

Auction Mechanism Design for Resources Allocation in Reservation Mode of Sharing Economy

共享经济给人们生活方式带来巨大改变,但是其预约模式大都采用固定价格、先预约先服务的方式分配资源,导致供应商收益低下,资源利用率也不高。基于此,提出一种基于拍卖的机制RAUPAM(resources allocation and user payment approximation mechanism),该机制适用于任何按时租赁收费的共享资源,如共享停车位、住宿等,通过拍卖来解决预约模式下共享资源的分配及用户定价问题。在资源分配问题上,该机制采用关键路径思想对每个资源进行全局规划,提高资源利用率的同时保证收益极大化;在价格支付问题上,基于临界值求解用户支付价格,尽可能地降低用户支付,吸引用户参与。实验结果表明,在不同的资源密度和分配周期下,RAUPAM都优于传统的FCFS(first-come-first-serve)和MAXBID算法,说明RAUPAM在预约高峰期和低峰期都表现更优,且不受分配周期的影响。通过详细对比,RAUPAM的利润比FCFS至少提高55%,比MAXBID算法平均提高30%;其服务用户率、资源利用率都比FCFS和MAXBID算法有显著提高。

The sharing economy has brought great changes to peoples lifestyle, but most of the reservation modes allocate resources by the method of fixed price and first-come-first-serve. This results in low profits of suppliers and low utilization of resources. Based on this, this paper proposes a mechanism based on auction, RAUPAM(resources allocation and user payment approximation mechanism), which is applicable to any sharing economy resource that is charged on time, such as shared parking spaces, accommodation, etc., and can solve the allocation of sharing resources and user pricing problem in the reservation mode based on auction. For the resource allocation problem,this mechanism adopts the idea of critical path to make global plan for each resource, which improves the resource utilization and guarantees the maximization of income. The payment algorithm takes into account critical value topayment as far as possible and attract users to participate. The experimental results show that RAUPAM is superior to the traditional FCFS(first-come-first-serve) and MAXBID algorithms in different resource densities and allocation cycles, which indicates that RAUPAM performs better in peak and low peak periods of reservation and is not affected by allocation cycles. Through detailed comparison, the profit of RAUPAM is at least 55% higher than FCFS and 30% higher than MAXBID;its service user rate and utilization of resources are significantly higher than FCFS and MAXBID.