列控系统RBC测试序列优化生成方法

Optimal generation method of RBC test sequence for train control system

目前,列控系统测试序列大多由人工编写而成,在测试项的有效性和测试案例覆盖的全面性等方面存在较多问题.针对传统蚁群算法收敛速度慢和易发生停滞现象等缺点,提出一种基于改进蚁群算法(Modified-Ant Colony Optimization,M-ACO)的测试序列优化生成方法,以RBC单电台切换场景为例,根据CTCS-3级列控系统技术规范构建RBC切换有色Petri网(Colored Petri Net,CPN)模型,由该模型生成状态空间可达图和可扩展标记语言(Extensible Markup Language,XML)文件,采用路径搜索算法生成满足全节点覆盖的测试案例集,根据各测试案例的开始条件和结束条件将测试案例串联后生成测试序列,再应用M-ACO算法生成优化测试序列.最后与序列优选算法及传统蚁群算法进行对比,结果表明:该方法降低了测试的复杂程度,测试利用率较未改进的蚁群算法提高了38.53%,适合复杂系统的测试.

Currently, most of the train control system test sequences are written manually, and there are many problems in the validity of test items and the comprehensiveness of test case coverage. Aiming at the shortcomings of traditional ant colony algorithm such as low convergence speed and stagnation, a test sequence optimization generation method based on Modified-Ant Colony Optimization(MACO) is proposed, taking the RBC single-station switching scenario as an example. According to the technical specifications of CTCS-3 train control system, an RBC-switched Colored Petri Net(CPN)model is built. Based on this model, state space reachability graph and XML(Extensible Markup Language) files are generated. The path search algorithm is used to generates a test case dataset that satisfies the full-node coverage, and the test cases are connected into a test sequence according to the start conditions and end conditions. Then, the M-ACO algorithm is used to generate optimized test sequence. Finally, the proposed method is compared with the sequence optimization algorithm and the ant colony algorithm. The results show that the method reduces the complexity of the test, and the test utilization rate is 38.53% higher than that of the original ant colony algorithm, which is suitable for the testing of complex systems.