资源敏感的软件自适应关键技术研究

The ever-increasing complexity of software systems and the change of runtime environment bring difficulties for modern software to run reliably and stably, making research on software self-adaptability of great theoretical and practical significance. The software resource contention in complex environment, distributed concurrent design, software resource dependency analysis and software intend understanding bring new challenges to software adaptability research. Based on this, the project will carry out in-depth research on self-adaptive technology on resource-sensitive software, based on statistical analysis and machine learning methods. The objectives of this project include: (1) automated analysis of resource usage based on program analysis techniques and formal methods; (2) automated mining of resource related API specification to provide clues for solving resource utilization issues; (3) resource dependency analysis and software intend understanding using program slicing techniques and dynamic testing methods; (4) monitoring software runtime status and resource usage in real time, and resolving resource contention problem; (5) building a runtime self-adaptive framework for long-term survival and evolution of software; (6) the formation of a software-adaptive solution targeted at the typical UAV field. By studying the above scientific issues, we believe this project can effectively enhance the software adaptability and ensure the software system running reliably and stably.

软件系统的日益复杂和软件运行环境的不断变化给软件系统的可靠稳定运行带来困难,使得软件自适应性研究具有重要的理论价值和现实意义.复杂环境下的多软件资源争用、软件的分布式并发设计、软件资源依赖分析和软件意图理解给软件自适应性研究带来了新的挑战.基于此,本课题将基于统计分析和机器学习等方法,研究资源敏感的软件自适应关键技术.具体包括(1)基于程序分析技术和形式化方法,研究资源使用不变式的自动分析技术;(2)自动化挖掘资源使用API的规约,为解决资源使用故障问题提供线索;(3)利用程序切片技术和动态测试方法,对软件进行资源依赖分析和软件意图理解;(4)实时监控软件运行时状态和资源使用情况,并针对资源争用进行冲突消解;(5)构建软件长期生存演化的运行时自适应支撑框架,通过在线和离线两种方式实现软件自适应调整;(6)在以上研究基础上针对无人机典型算法,形成面向无人机领域的软件自适应解决方案.

本项目针对复杂软件环境下资源敏感的软件自适应关键技术展开研究。针对复杂环境下的多软件资源争用、软件的分布式并发设计、软件资源依赖分析和软件意图理解给软件自适应性研究带来的挑战，基于程序分析技术挖掘资源使用API的规约,为解决资源使用故障问题提供线索;利用程序切片等技术对软件进行资源依赖分析和软件意图理解;通过实时监控软件运行时状态和资源使用情况,并针对资源争用进行冲突消解;构建软件长期生存演化的运行时自适应支撑框架,实现软件自适应调整;并针对无人机典型应有场景,形成面向无人机领域的软件自适应解决方案。课题组按照研究计划推进研究工作，完成了课题拟定目标，在资源依赖分析、软件意图理解、运行时自适应支撑框架等方面取得了多个关键技术突破。截止2022年底课题组共在国内外学术期刊、学术会议上发表学术论文12篇，其中SCI检索4篇、EI检索7篇。申请专利2项。培养博士研究生3人，硕士研究生5人。