SaaS软件的部署对其性能有着重要的影响。如何自适应地优化SaaS软件的部署，改善服务性能，是一个具有重要学术意义和应用价值的问题。本项目将研究面向性能改善的SaaS软件部署方案自优化方法。具体包括：针对SaaS软件性能评估和优化的需求，设计一种SaaS软件部署描述语言SSDL；基于分层排队网络，设计一种SaaS软件部署方案的性能模型，并基于加权平均方法和分层方法给出部署方案的性能评估方法；基于粒子动力学模型和统计物理理论，设计SaaS软件部署方案的优化算法；提出SaaS软件部署方案的动态自优化方法，并研发其支撑工具，支持SaaS软件的工程化实践。通过本项目的研究，揭示SaaS软件部署对性能的影响机理，形成性能驱动的SaaS软件部署自优化的理论和方法。

Deployment of SaaS (Software-as-a-Service) software has significant impacts on its performance. It is of both high academic and practical value to optimize the deployment of SaaS software in order to improve the service performance. This project will explore a self-optimizing approach for SaaS software deployment toward performance improvement. Specifically, several works will be done for the deployment of SaaS software. A description language SSDL is to be presented to meet the need of performance evaluation and optimization for SaaS software; a performance model is to be proposed based on layered queuing network, and furthermore a performance evaluation method was given based on the model with WAM (weighted average method) and MOL (method of layers); an optimization algorithm is to be designed based on the particle dynamics model and statistical physics theory; a self-optimizing method is to be proposed and its supporting tools are to be developed in order to support the engineering practice of SaaS software. In this project, the mechanism of how deployment of SaaS software influences its performance will be revealed, which will further form the theory and method of performance-driven self-optimization of SaaS software deployment.