含有不确定性的复杂系统高效仿真及其评估方法

Uncertainty is an important characteristic of complex systems, and modeling & simulation(M&S) would also introduce uncertainties. To meet the simulation and assessment requirements of complex system with numerous uncertaities, firstly,.uncertaity will be analyzed based on M&S development process model, and the classification method and mathimetical description of uncertainty will be presented; Secondly, high-efficiency simulation methods of complex system with uncertainties will be studied, including nested layered sampling method, the method based on sensitivity analysis and experiment design, and the method based on meta model; Followingly, simulation assessment methods considering the influence of uncertainty will be proposed, including a multi-output simulation validation method, and a method based on probability-box to quantify the simulation output total uncertaity; Finally, the proposed methods and corresponding tool of this research project will be applied in an actual complex simulation system for the validation purpose. The research results could heighten the complex simulation experiment's efficiency and enrich the simulation assessment methods, meanwhile promote the application of complex system simulations..

不确定性是复杂系统的重要特征，同时建模与仿真也会引入不确定性。本项目针对含有众多不确定性因素的复杂系统仿真及评估需求，首先基于仿真开发过程模型分析仿真中各种类型和来源的不确定性，给出相应的分类及数学描述方法；然后，为了提高仿真效率，研究给出含有不确定性的复杂系统高效仿真方法，包括嵌套的分层抽样方法、基于灵敏度分析和实验设计的方法以及基于元模型的方法；进而提出考虑不确定性影响的仿真评估方法，包括考虑不确定性的多元仿真输出验证方法及基于概率包络的仿真输出整体不确定性量化方法；最后将项目研究成果在实际复杂仿真系统中进行应用与验证。本项目的研究成果具有重要的理论意义与实用价值,可提高复杂仿真实验效率、丰富仿真评估方法体系,推动复杂系统仿真的应用深度与广度。

针对含有众多不确定性因素的复杂仿真系统高效实验及评估需求，完成了仿真不确定性的分类及描述方法、含有不确定性的复杂仿真系统高效实验方法、多元不确定性仿真结果评估验证方法及复杂系统高效仿真及其评估应用等研究内容，建立了仿真不确定性基础理论框架，提出了基于证据理论的复杂仿真结果验证方法，从提高仿真抽样效率、降低不确定性因素维度、模型替代、增量扩展等视角分别给出了高效仿真实验方法，进而开发了复杂系统高效仿真及评估辅助工具软件。上述成果具有重要的理论意义与实用价值，可提高复杂仿真实验效率、丰富仿真可信度评估方法体系。