Software Development for Prediction of Risk in the Transport Infrastructure.

Within this final qualification work, a software complex was developed for predicting risks in the transport infrastructure, which collects the statistical data from the open sources and predicts the time delays of the flights. The electronic online board of arrivals and departures of the Domodedovo airport aircraft has been selected as a source of data. The complex was designed using the software products that have a universal public license and are distributed free of charge to the developers. The development was conducted using a high-level general-purpose programming language Python. As a hardware-computing infrastructure, a virtualized infrastructure was used based on the Oracle VM VirtualBox software and the Linux OpenSUSE operating system. When designing the complex and selecting software tools for its implementation, the requirements stated in the terms of reference are taken into account. The selected components of the infrastructure fully comply with the principle of cross-platform. When choosing a model of the relational database as the main data repository, such criteria as linear scalability and fault tolerance are fully taken into account and implemented. The DBMS is optimized to the most frequent requests, and provides a minimum response time, as well as guarantees a reliable and predictable work. The modeling of risk assessment is based on the Monte Carlo method. The mathematically determined numerical method showed an estimated level of significance of the result in 93%, with the number of iterations equal to 1000. Modeling and solving practical problems confirmed theoretical values and showed that the developed method allows solving problems with a given accuracy for acceptable computational time. All software modules passed the life cycle of software development: design, creation, testing. Fig.:8. Tab.: 3. Ref.: 17 titles.