面对高复杂软件测试所面临的测试目标多样化，测试复杂度高等特性，多目标优化算法的快速收敛、三维或者更高维优化目标函数的科学化构建、种群多样性机制的合理设计等关键问题的解决变得极具挑战性。本项目拟通过对多目标测试用例选择指标体系、基于Pareto最优的多目标测试用例约简框架的搭建，群体智能搜索算法中种群多样性机制设计等几方面主要内容的研究。结合机器学习技术、统计理论、以及最优化技术等的综合应用，研制具多样性并融合Pareto最优的多目标优化软件测试用例约简群体智能搜索算法，力求突破高复杂软件测试中的一些瓶颈问题，对推动最优化理论、机器学习理论和群体智能搜索技术在新一代信息技术领域的应用和研究具有重要科学意义。

Given the diversification of test objectives, higher complexity of software testing, it has become more challenging to solve such problems as the fast convergence of multi-objective optimization algorithm, the construction of three-dimensional or multidimensional optimized objective functions and the design of population diversity mechanism. The project will maily study the following aspects. The first is the choice in indicator system of multi-objective test case reduction. The second is the construction of multi-objective test case reduction framework based on Pareto optimization. And the last is the design of population diversity mechanism in group intelligent search algorithms. With the intergrated application of optimization technology, statistical theory and machine learning technology, the research aims to develop a multi-objective optimization software test case reduction model based on group intelligent search algorithm with diversity, and hybrid with Pareto optimization. It is of great significance to promote the research and application of optimization theory, machine learning theory and group intelligent search technology in the new generation of information technology.