面向不确定环境的混成AADL设计与量化分析

发布时间：2018-03-23 04:13

本文选题：AADL　切入点：不确定环境　出处：《华东师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：混成AADL由于其简单的语法、强大的功能以及良好的扩展性等特点,常被用作信息物理融合系统的体系架构模型设计。其设计模型既要能够描述信息物理融合系统中的混成行为,又要作为"黄金参考模型"支持底层实现的产生,因此其正确性与可靠性是系统构造的关键。然而,混合AADL设计模型常采用的最差情况分析方法往往导致过度悲观的分析结果,并且当系统与不确定环境紧密相互作用时,这种方法无法对系统的整体性能准确分析。针对混成AADL在不确定环境下的建模与分析问题,本文提出了面向不确定环境的混成AADL设计与量化分析方法。该方法能够有效的支持混成AADL的不确定性建模,同时支持对混成AADL模型的定量分析。本文主要贡献如下:1.AADL不确定性附属语言针对信息物理融合系统所处的不确定环境,本文提出了一种AADL不确定性附属语言。不确定性附属语言不仅可以精确的对系统交互环境引起的性能变化建模,还可以方便的定义系统设计者的需求规约,有效扩充了 AADL的不确定性建模方法。2.不确定环境下混成AADL设计量化分析框架本文提出了一种不确定环境下混成AADL系统设计的量化分析框架。本框架将不确定环境下的混成AADL系统模型转换为价格时间自动机模型,通过统计模型检验对转换后的价格时间自动机验证,支持对混成AADL模型在不确定环境下的量化分析。3.自动化的模型转换工具AADL2UPPAAL本文设计并实现了自动化的模型转换工具AADL2UPPAAL,并将其集成到量化工具链中。该量化工具链包括AADL建模、模型转换和量化分析三大功能,具备完整的量化分析AADL模型的能力。通过本文提出的工具链可以快速、准确的比较不同AADL设计方案的性能差异。中国列车控制系统3级中的行车许可场景实验表明,本文提出的不确定性附属语言能够有效地对不确定环境建模,量化分析方法能够帮助AADL设计者定量分析和选取设计模型,同时,自动化的模型转换工具能够准确且快速的支持量化分析。
[Abstract]:Because of its simple syntax, powerful function and good expansibility, hybrid AADL is often used in architecture model design of information physics fusion system, whose design model should be able to describe the mixing behavior in information physics fusion system. As a "gold reference model", it is necessary to support the generation of the underlying implementation, so its correctness and reliability are the key to the system construction. However, the worst-case analysis method often used in hybrid AADL design models often leads to overly pessimistic analysis results. When the system interacts closely with the uncertain environment, this method can not accurately analyze the overall performance of the system. In this paper, a hybrid AADL design and quantitative analysis method for uncertain environments is proposed, which can effectively support the uncertainty modeling of mixed AADL. The main contributions of this paper are as follows: 1. The uncertain adjunct language of AADL is based on the uncertain environment of information physics fusion system. In this paper, a AADL uncertainty dependent language is proposed, which can not only accurately model the performance change caused by the system interaction environment, but also define the requirement specification of the system designer. This paper extends the uncertainty modeling method of AADL effectively .2.The quantitative analysis framework of mixed AADL design in uncertain environment is presented in this paper. A quantitative analysis framework for hybrid AADL system design in uncertain environment is presented. The framework will be uncertain ring. The hybrid AADL system model is transformed into the price time automata model. The price time automata after conversion are verified by statistical model test. Support for quantitative Analysis of mixed AADL Model in uncertain Environment .3.Automated model transformation tool AADL2UPPAAL this paper designs and implements an automated model transformation tool AADL2UPPAAL. and integrates it into the quantization tool chain. The quantization tool chain includes AADL modeling. The function of model transformation and quantitative analysis has the capability of quantifying AADL model. The tool chain proposed in this paper can be used quickly. The performance differences of different AADL design schemes are compared accurately. The results of train permission scenarios in China's train control system level 3 show that the uncertainty dependent language proposed in this paper can effectively model the uncertain environment. The quantitative analysis method can help the AADL designers to quantitatively analyze and select the design model. At the same time, the automated model transformation tool can support the quantitative analysis accurately and quickly.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP29

【参考文献】