Reliability Modeling And Semi-physical Simulation Of Interlocking Software Based On Non-homogeneous Poisson Process

Computer interlocking system is an important part in railway signal system.Its reliability is the important premise to ensure the safe and efficient operation of trains.With the improvement of the hardware structure of interlocking system,the reliability of the system is more and more dependent on interlocking software.In the lifecycle of software,test phase is an important part for improving its reliability.Using the test failure data to establish reliability growth model thus to further evaluate and predict the growth of reliability is an important research direction for the reliability field of software.There are nearly hundred kinds of proposed reliability growth models,but if these ordinary models are used to safety critical software like interlocking software,the reliability evaluation and prediction performance will be defective.In order to establish the reliability growth model which is more suitable for interlocking software,the following research is carried out in the paper.Firstly,the background knowledge of interlocking system and software reliability engineering is introduced.The reliability model of Non-homogeneous Poisson Process(NHPP)which is used widely is selected as the mathematical foundation of modeling,and moreover,few kinds of idea of modeling for classical NHPP reliability models are researched and analyzed.Secondly,the test process of interlocking software is analyzed.It is found out that there are incomplete testing coverage and imperfect debugging during software test and fault correcting.Considering these phenomena into modeling assumptions,the differential equation is established and the appropriate expression for the special function is selected.Therefore,an interlocking software reliability growth model is established.Then,a set of actual testing interlocking software data is used to verify this model,and results show that comparing with several classical software reliability growth models,the improved model can better evaluate and predict the reliability growth situation of interlocking software.Finally,according to the Technological Requirements of Computer Interlocking,the computer interlocking software which can receive operation orders,judge interlocking logics,and accurately perform is designed and realized in the paper.Moreover,the corresponding hardware platform is constructed to coordinate it to fulfill the complete interlocking functions,the functional test of this interlocking software is implemented with test cases,thus,software failure phenomenon and failure data are recorded.In the later period of software test,the interlocking software reliability growth model proposed in the paper is used to evaluate the reliability of the software,and moreover,the time for software reaching the expected reliability standards is predicted.The interlocking software reliability growth model is established in the paper.The evaluation and prediction effects of the model on the software reliability are verified through the actual software failure data.Interlocking semi-physical simulation system is designed and realized in the paper,and moreover,the model proposed in the paper is used to the software test phase of this system.Results show that the model can accurately evaluate the reliability development of the software and predict the time for software reaching the expected reliability standards,and moreover,the evaluation and prediction results will provide the important scientific basis for the development management of computer interlocking system.