基于分割法的混杂系统形式化分析方法研究

发布时间：2018-07-11 12:33

本文选题：混杂系统 + 近似分析　；参考：《广西民族大学》2017年硕士论文

【摘要】：混杂系统是连续变量动态系统和离散事件动态系统相结合,并且相互交换信息的一类系统。其中连续子系统的动态特征随时间的发展而不断变化,离散性子系统的动态演化受事件的驱动,二者相互作用,使系统的运行轨迹整体上呈现离散位置的迁移,局部上呈现连续状态的渐进演化,在综合了连续变量动态系统和离散时间系统演化特征的基础上,表现出更加复杂的动态行为。混杂系统的形式验证是混杂系统的重要研究方向之一,它在给定的初始条件下,分析系统的可达集是否都在目标状态集合内。本文围绕混杂系统的近似分析方法展开研究,从混杂系统本身的特点出发,以求解混杂系统的可达集为目标,并用多面体表示可达集,具有良好的保守性,但是多面体的计算非常复杂,且随着维数增长,多面体的顶点和面的个数增长迅速,在系统维数高时很难处理。所以求解出一种新的计算近似多面体体积的方法----分割法。该方法是依据分割法的原理,结合混杂自动机的定义,构造一个基于分割的近似混杂自动机进而求解可达集;然后分析基于分割的近似混杂自动机的行为等价。
[Abstract]:Hybrid system is a kind of system which combines continuous variable dynamic system with discrete event dynamic system and exchanges information with each other. The dynamic characteristics of the continuous subsystem change with the development of time, and the dynamic evolution of the discrete subsystem is driven by events. On the basis of integrating the evolution characteristics of dynamic systems with continuous variables and discrete-time systems, it shows more complex dynamic behavior. Formal verification of hybrid systems is one of the important research directions of hybrid systems. Under given initial conditions, the reachability sets of hybrid systems are all in the target state set. In this paper, the approximate analysis method of hybrid systems is studied. Based on the characteristics of hybrid systems, the objective is to solve the reachability sets of hybrid systems, and the reachability sets are represented by polyhedrons. But the calculation of polyhedron is very complicated, and with the increase of dimension, the number of vertex and surface of polyhedron increases rapidly, which is difficult to deal with when the dimension of the system is high. Therefore, a new method to calculate approximate polyhedron volume-segmentation method is proposed. According to the principle of the segmentation method and the definition of hybrid automata, this method constructs an approximate hybrid automaton based on segmentation to solve the reachable set, and then analyzes the behavior equivalence of the approximate hybrid automata based on segmentation.
【学位授予单位】：广西民族大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O175

【参考文献】