信息不对称条件下针对蓄意攻击的军用微电网最优防御技术研究

Due to the wide integration of information technology, stable and reliable power supply has become one of the pivotal factors in modern warfare to achieve victory. On the other hand, according to Warden's Five Rings Theory, military microgrid (MMG) will be the primary attack target, and the defense of MMG has gained significant attention in the community of both military and academic. However, the existing research has the following limitations: 1) the information symmetry between offensive and defensive sides is not considered, resulting in inaccurate mathematical model and low defense efficiency; 2) the timeliness of MMG defense process is lack of concern, leading to a passive situation as the defensive plan might not be updated timely. In order to overcome the model defects and solve the practical problems, this project takes MMG defense problem as the main topic, follows a logical process of "modeling - optimization - acceleration", and implements the investigation of optimal fortification techniques for MMG against deliberate attack under information asymmetric conditions: 1) the information asymmetry between two opposite roles is introduced in the defensive model for the first time; 2) an efficient and effective algorithm framework is proposed based on the adaptive robust optimization algorithm; 3) the convergent property and execution efficiency of the developed algorithm are improved based on optimization theory and GPU parallel computing. This project is not only beneficial in addressing practical problems but also stimulates the development of related fields, including information asymmetric binary game, deterministic optimization algorithm, and parallel computing.

由于信息技术的广泛嵌入，稳定可靠的电能供应和保障已成为决定战争胜负的关键因素之一。另外，根据“五环”目标理论，军用微电网将是敌方重点打击对象，因此军用微电网防御已成为一个值得研究的关键问题。然而，现有研究存在以下局限：1）对攻防双方的信息不对称性缺乏考虑，导致模型精度不够且防御效率低下；2）对微电网防御过程的时效性缺乏认识，导致决策方案更新不及时而陷入被动。为了弥补模型缺陷并解决实际问题，本项目以军用微电网防御问题为研究对象，以“系统建模–优化算法–加速改进”为逻辑导向，开展信息不对称条件下针对蓄意攻击的军用微电网最优防御技术研究，首次将攻防双方信息不对称特性引入模型构建过程，并基于自适应鲁棒优化框架给出科学有效的求解算法，最后利用优化理论和GPU并行计算改进算法的收敛特性和执行效率。项目研究成果不仅用于解决实际问题，还将促进信息不对称条件下二元博弈、精确优化算法和并行计算等领域的发展。

项目分析研究了微电网最优防御面临的建模与优化问题和特点，着重考虑了战场条件下信息不对称带来的影响，将攻防双方信息不对称特性引入模型框架构建过程并给出科学有效的求解算法，此外基于优化理论和启发式策略改进算法的收敛特性和执行效率，使得整个“建模-优化-求解”框架符合实际，确保战争对抗条件下军用微电网能够发挥应有效能，为赢得战斗胜利创造条件。.首先，项目研究成果提出的微电网攻防对抗信息不对称评价指标，大大提高了微电网攻防博弈模型的真实性，增强了攻击和防御决策对实际过程的指导意义，同时在该研究的引领下，后续微电网攻防博弈研究均可考虑双方信息不对称特性的影响，促进相关领域研究的进步和发展。.其次，项目研究成果构建的微电网攻防模型加入了信息不对称的影响，并考虑隐藏策略、欺骗策略、保护策略、物理攻击、信息攻击、协同攻击、防御资源最优部署和调度等多种因素，形成了较为完整的微电网攻防博弈定量分析框架，为未来改进提供参考原型。.再次，项目研究成果提出的多种鲁棒优化求解算法，包括改进型CCG、嵌套型CCG、多目标鲁棒优化、期望鲁棒优化等算法和策略，讨论了每种算法的精确性、收敛性、方案保守性、应用场景等，形成了求解鲁棒优化典型问题的工具库，为仿真平台的搭建奠定了重要基础。.最后，项目研究成果提出的优化算法与数值计算加速策略，包括分级优化或离散化近似等方法，可针对特殊典型问题和场景，在保证工程应用精度要求的前提下，大大提高原始算法的求解效率，增强决策方案的时效性，为决策者求解大规模问题提供了可行方案。