基于融合聚类的新型列控系统测试用例自适应排序方法  

Adaptive Sorting Method for Test Cases of New Train Control System Based on Fusion Clustering

作　　者：张素阳 窦道飞 郝韬 ZHANG Suyang;DOU Daofei;HAO Tao(Beijing Huatie Information Technology Co.,Ltd.,Beijing 100081,China;Signal&Communication Research Institute,China Academy of Railway Sciences Corporation Limited,Beijing 100081,China;Railway Industry Engineering Research Center for Intelligent Operation and Maintenance of Communication&Signaling Infrastructure,Beijing 100081,China)

机构地区：[1]北京华铁信息技术有限公司,北京100081 [2]中国铁道科学研究院集团有限公司通信信号研究所,北京100081 [3]通信信号基础设施智能运维铁路行业工程研究中心,北京100081

出　　处：《中国铁路》2025年第2期28-34,共7页China Railway

基　　金：中国铁道科学研究院集团有限公司科研开发基金项目(2024YJ331)。

摘　　要：为提高新型列控系统功能故障检测能力,缩短测试周期,提出聚类分析与多目标优化相结合的测试用例排序方法。分析新型列控系统结构及功能,提取临时限速场景中各子系统的关键状态,生成初始测试用例集;进行文本聚类和覆盖聚类分析,将聚类结果按一定的权重组合,得到测试用例相似性矩阵;以平均故障检测率、语句覆盖率和有效执行时间为多目标优化的度量指标,采用可支配遗传算法与历史执行信息对测试用例集进行自适应排序,得到最终测试序列。依托京张高铁列控系统仿真测试平台,验证该方法的有效性。试验结果表明:在不同测试用例规模情况下,该方法与AP、AC-CAP、AAR-CAP方法相比较,标准平均故障检测率的表现更优,排序结果具有更好的故障检测效果,可为新型列控系统功能测试提供新的、高效的解决方案。The paper proposes a test case sorting method that combines cluster analysis and multi-objective optimization to enhance the fault detection capability of the new train control system and shorten the test cycle.It analyzes the structure and functions of the new train control system and extracts the key states of each subsystem in temporary speed restriction scenarios to generate an initial test case set.The paper then studies text and coverage clusterings and pools the clustering results using specific weights to create a test case similarity matrix.The average fault detection rate,statement coverage,and effective execution time are used as multi-objective optimization metrics.An adaptive sorting of the test case set is conducted using a dominance-based genetic algorithm and historical execution information,resulting in the final test sequence.The effectiveness of this method is validated using the simulation test platform for the train control system of Beijing-Zhangjiakou High Speed Railway.The test results indicate that,under varying test case scales,the method outperforms the AP,AC-CAP,and AAR-CAP methods in terms of standard average fault detection rate.The sorting results demonstrate improved fault detection effectiveness,providing a new and efficient solution to the functional testing of the new train control system.

关 键 词：新型列控系统 测试用例排序 聚类分析 组合测试 临时限速 故障检测 

分 类 号：U284.48[交通运输工程—交通信息工程及控制]