基于深度学习与遗传算法的动车组与ATP车载设备接口试验测试序列优化生成

Optimized Generation of Test Sequences for Interface Type Test between Train and ATP Onboard Equipment Using Deep Learning and Genetic Algorithm

接口型式试验验证不同类型ATP车载设备与不同型号动车组之间接口的适配性与安全性,而其测试序列设计和优化对在有限试验资源下高效、安全地完成试验十分重要。目前,接口型式试验的测试序列主要基于经验人工编写形成,测试项的有效性和测试序列的优化衡量等方面存在较多问题。本文研究了接口型式试验的测试序列的优化生成方法,利用深度学习算法与遗传算法,针对测试序列的生成与优化进行建模,设计测试序列的生成与优化两个相互联系的过程,综合地在序列解空间中寻找次优解,以得到期望的测试序列。基于型式试验现场数据进行了仿真与分析,论述了本文策略的有效性。

Interface type test verifies the suitability and safety of interface between different types of EMU trains and different ATP onboard equipment.The design and optimization of the test sequences are significant for the efficient and safe completion of the experiment under limited test resources.The efficiency of conducting interface type test cannot be optimal since test sequences of the interface type tests are now manually generated according to test experience.This paper studied the optimization method of generating test sequences of the interface type tests and proposed a strategy to generate optimized test sequences for interface type tests,which integrated deep learning based test sequence updating process with genetic algorithm based test sequence optimizing process.Models of generating,updating and optimizing test sequences were built to seek sub-optimal solution for the expected optimized test sequences depending on the comprehensive optimized objectives.Based on the simulations and analysis of the field data of type test,the proposed strategy was proved to be effective to generate expected optimized test sequences.