基于多租户的高性能地震保险损失评估SaaS云平台设计与实现

地震巨灾风险的特点是低频高损，历史震害数据缺乏、风险暴露快速变迁等因素导致基于大数定理的费率厘定方法无法针对各区域不同建筑类型的风险暴露进行精细化定价。本文基于“五代图”潜在震源区模型的随机事件集解决观测数据不足的问题；并使用“五代图”所采用的地震动参数衰减关系模型与工程易损性方法计算地震事件对风险暴露造成的损失，从而计算费率厘定、地震风险管理需要的必备参数。本文采用云计算平台的弹性伸缩计算技术，实现动态按需分配计算资源，满足多用户并发使用的业务需求；同时采用以业务数据为单元的数据隔离方案，构建支持多租户的高性能地震保险损失评估SaaS云平台。

Design and Implementation of Multi-tenant SaaS Cloud Platform for High-Performance Seismic Insurance Loss Assessment

Earthquake catastrophe risk is characterized by low frequency and extreme severity. Due to the lack of historical earthquake loss data and the rapid changes of exposure, it is difficult to price the various exposure in different regions precisely based on the law of large numbers. And in this paper, the problem of insufficient data is solved by the application of the random event set based on the potential seismic source zones of the Fifth Generation National Seismic Zoning Map. The losses caused by those earthquakes are calculated by utilizing the Ground Motion Prediction Equation (GMPE) adopted by the Fifth Generation National Seismic Zoning Map and classic engineering vulnerability methods. Based on the output, we calculate the necessary parameters for insurance pricing and risk managing of seismic catastrophe. To meet the concurrent use needs of a large number of users, we implement the dynamic on-demand management of computing resources by adopting the elastic scaling computing technology based on the cloud computing platform. And to ensure the separation and security of user data, we isolate and split the data by the exposures uploaded by users. By the methods above, we implement a multi-tenant high-performance seismic insurance loss assessment SaaS cloud platform.