多目标优化算法及其在航电健康管理系统中应用

发布时间：2018-06-22 19:39

本文选题：多目标优化问题 + 测试选择问题　；参考：《电子科技大学》2017年硕士论文

【摘要】：由于新一代飞机对先进航空电子系统的需求越来越高,航空电子技术在近年得到了迅速发展。在航空电子系统功能日益完善的同时,其可靠性也得到越来越多研发人员的重视。航电健康管理系统作为航空电子系统可靠性的保障,逐渐成为研究热点,尤其是作为航电健康管理系统基础的测试选择问题。经典的测试选择方法是单目标方法,而实际上测试选择问题是典型的多目标问题。本文将多目标优化方法应用到航电健康管理设计的测试选择问题中,完成了以下研究工作:(1)完成了多目标优化方法研究。多目标优化是当前优化和决策领域的热点方向,它适用于存在多个需要同时考虑的优化目标并且这些优化目标往往互相冲突的情况。解决这类问题需要特殊的建模方式、优化理论和求解算法。因此,本文在具体研究了多目标优化问题的数学模型和Pareto最优理论的基础上,详细分析了MOGA、NSGA2、SPEA2、PAES这四种在多目标优化领域具有代表性,评价较高的算法,同时研究了多目标优化算法的性能评价方法。(2)详细研究了测试选择问题,分析了其中多个优化目标的物理意义和计算方式。介绍了基于故障字典技术的测试选择方法,研究了从初始故障字典构建,模糊组划分到整数编码转换的整个建模流程,还研究了贪婪选择法和智能优化法两类测试选择方法并分析了其优缺点。同时提出了一种基于测试选择优化的故障诊断技术。(3)提出了基于混沌多目标粒子群优化算法的测试选择方法。在离散粒子群算法的基础上进行了多目标优化的改进,特别加入混沌变异机制提高算法的全局搜索能力。实验证明混沌机制能避免算法出现早熟收敛现象,随后通过与其它算法的对比实验,验证了本文所提出的算法的有效性和卓越性。(4)完成了测试选择及故障诊断软件的设计与验证。结合本文提出的针对测试选择问题的多目标优化算法以及故障诊断技术,设计了一款由测试选择模块、实时监测模块和故障诊断模块组成的软件,并且通过该软件,验证了测试选择以及多目标优化方法应用的有效性。本文是多目标优化方法在航电健康管理系统中应用的尝试,为后续工作提供基础和新的研究方向。
[Abstract]:Avionics technology has been developed rapidly in recent years due to the increasing demand for advanced avionics systems for new generation aircraft. As the function of avionics system becomes more and more perfect, more and more researchers pay attention to its reliability. Avionics health management system, as the guarantee of avionics system reliability, has gradually become a research hotspot, especially the test selection as the basis of avionics health management system. The classical test selection method is a single objective method, but in fact the test selection problem is a typical multi-objective problem. In this paper, the multi-objective optimization method is applied to the test selection problem of avionics health management design. The following research work is accomplished: (1) the multi-objective optimization method is studied. Multi-objective optimization is a hot topic in the field of optimization and decision-making. It is suitable for the situation where there are many optimization objectives that need to be considered simultaneously and these optimization objectives often conflict with each other. Solving this kind of problems requires special modeling, optimization theory and solving algorithm. Therefore, on the basis of studying the mathematical model and Pareto optimal theory of multi-objective optimization problem in detail, this paper analyzes in detail four algorithms which are representative and highly evaluated in the field of multi-objective optimization. At the same time, the performance evaluation method of multi-objective optimization algorithm is studied. (2) the test selection problem is studied in detail, and the physical meaning and calculation method of multiple optimization objectives are analyzed. This paper introduces the test selection method based on fault dictionary technology, and studies the whole modeling process from initial fault dictionary construction, fuzzy group partition to integer coding conversion. Two kinds of test selection methods, greedy selection method and intelligent optimization method, are also studied and their advantages and disadvantages are analyzed. At the same time, a fault diagnosis technique based on test selection optimization is proposed. (3) A test selection method based on chaotic multi-objective particle swarm optimization algorithm is proposed. Based on the discrete Particle Swarm Optimization (DPSO), the multi-objective optimization is improved, especially the chaos mutation mechanism is added to improve the global search ability of the algorithm. The experimental results show that the chaotic mechanism can avoid premature convergence of the algorithm, and then compared with other algorithms, The validity and excellence of the proposed algorithm are verified. (4) the design and verification of test selection and fault diagnosis software are completed. Combined with the multi-objective optimization algorithm and fault diagnosis technology proposed in this paper, a software composed of test selection module, real-time monitoring module and fault diagnosis module is designed. The validity of test selection and application of multi-objective optimization method is verified. This paper is an attempt to apply multi-objective optimization method in avionics health management system, which provides the basis and new research direction for the follow-up work.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V243;O221.6

【参考文献】