首页> 外文会议>Formal Methods in Computer-Aided Design, 2010 >Incremental component-based construction and verification using invariants
【24h】

Incremental component-based construction and verification using invariants

机译:基于组件的增量式构造和使用不变量的验证

获取原文

摘要

We propose invariant-based techniques for the efficient verification of safety and deadlock properties of concurrent systems. We assume that components and component interactions are described within the BIP framework, a tool for component-based design. We build on a compositional methodology in which the invariant is obtained by combining the invariants of the individual components with an interaction invariant that takes concurrency and interaction between components into account. In this paper, we propose new efficient techniques for computing interaction invariants. This is achieved in several steps. First, we propose a formalization of incremental component-based design. Then we suggest sufficient conditions that ensure the preservation of invariants through the introduction of new interactions. For cases in which these conditions are not satisfied, we propose methods for generation of new invariants in an incremental manner. The reuse of existing invariants reduces considerably the verification effort. Our techniques have been implemented in the D-Finder toolset. Among the experiments conducted, we have been capable of verifying properties and deadlock-freedom of DALA, an autonomous robot whose behaviors in the functional level are described with 500000 lines of C Code. This experiment, which is conducted with industrial partners, is far beyond the scope of existing academic tools such as NuSMV or SPIN.
机译:我们提出基于不变式的技术,以有效验证并发系统的安全性和死锁属性。我们假设组件和组件之间的交互是在BIP框架(基于组件的设计工具)中描述的。我们基于一种组合方法,在该方法中,通过将各个组件的不变量与考虑了组件之间的并发和交互作用的交互不变量相结合来获得不变量。在本文中,我们提出了计算交互不变性的新有效技术。这可以通过几个步骤来实现。首先,我们提议基于组件的增量设计的形式化。然后,我们提出了充分的条件,以通过引入新的交互作用来确保不变量的保留。对于不满足这些条件的情况,我们提出了以增量方式生成新不变量的方法。现有不变量的重用大大减少了验证工作。我们的技术已在D-Finder工具集中实现。在进行的实验中,我们已经能够验证DALA的属性和无死锁,该机器人的功能级别上的行为已用50万行C代码描述了自主机器人。与行业合作伙伴进行的这项实验远远超出了现有的学术工具(如NuSMV或SPIN)的范围。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号