An efficient method of disturbance analysis and train rescheduling using MWDLNN and BMW-SSO algorithms

The ability to efficiently reschedule the trains when disturbances occur is one among the vital factors in attaining higher punctuality on railway traffic systems. Numerous techniques were developed to carry out train rescheduling, but the existing methodologies don't consider the probability of conflicts (or) disturbance analysis (DA). Thus, in this paper, an efficient method of train rescheduling (TRS) system is proposed utilizing the Brownian motion weighted-based salp swarm optimization (BMW-SSO) algorithm. The proposed technique follows two processes: (i) DA and (ii) rescheduling. Primarily, the track detail is collected as an input, and the disturbances are identified utilizing the modified weight-based deep learning neural network. If any disturbance occurs, then the TRS process is performed utilizing the BMW-SSO. Aimed at this, the current timetable is utilized as an input, and some constraints are extracted from the input utilizing multi-choice mixed integer goal programming. After that, finally, the BMW-SSO chooses and optimizes the best constraints from the extracted constraints as well as reschedules the timetable. Here, the BMW-SSO alternatively chooses the best third rescheduling timetable. Subsequent to generating TRS, the first rescheduled timetable's feasibility is checked centred on the dictionary-centred checking technique. If the first rescheduled timetable is feasible, then it is signified as the optimal timetable. Otherwise, the second-created rescheduled timetable is utilized. And if the second one is unfeasible, then the thirdly created timetable is utilized. At last, the experiential examination confirms the proposed train timetable rescheduling (TTR) system's performance regarding the timetable deviation, operation cost, along with dwell time. The proposed TTR consumes low time while analogized with the scheduled table.