While determining the travel time between stations, a number of design parameters such as waiting time, motion resistance, slope, curve, traction force, maximum speed, vehicle mass, and distance between two stations are taken into consideration. These parameters form the infrastructure of the system definition of the motion of the vehicle. Furthermore, while creating the speed profile, special attention should be paid to the travel time in order to ensure the defined headway for the line. In this study, the travel time value between stations for intracity metro stations was predicted using the adaptive boosting method, which is one of the machine learning methods, and compared with various well-known methods. The data used were applied to the proposed model with the cross-validation and random sampling hold-out methods, and the values of the coefficient of determination (R2) were calculated.
Transportation systems take an essential place in traffic planning. While designing railways, transportation between access points is planned to be realized within the desired time. The average speed between access points is affected by parameters like waiting time, motion resistance, slope, curve, traction force, maximum speed, the mass of the vehicle, and distance between two stations. The motion of the vehicle is calculated with these parameters, and the system design is performed accordingly. The average speed is one of the most critical factors affecting travel time between two access points. The headway may vary depending on the average speed. In this study, different regression methods were applied to estimate the average speeds calculated between stations in rail systems, and the obtained successful results were presented comparatively. When the methods used were examined, the Gaussian process regression method, which was optimized with the Bayesian algorithm, was observed to yield the most successful result. A Gaussian process (GP) is a collection of random variables, any finite number of which have a Gaussian distribution. Following simulations, the root mean square error and mean absolute error were found to be 0.064 and 0.047, respectively, and the coefficient of determination (R2) value was obtained as 1 when the success rate of the method was calculated.
22 January 2019 in Turkish Journal of Electrical Engineering and Computer Sciences
Modeling the operational dynamics of intricate rail transit systems faces three significant challenges: addressing network-wide dependencies, differentiating correlation from causation, and accurately quantifying prediction uncertainty. Current methodologies generally tackle these issues separately, resulting in models that are either structurally simplistic, causally unclear, or overly confident in their forecasts. This paper presents a comprehensive framework that, for the first time, effectively combines Graph Neural Networks (GNNs), Causal Machine Learning (CML), and Conformal Prediction (CP) to resolve this dilemma. GNNs are utilized to capture the topological dependencies within the rail network, CML is employed to discern the unbiased causal impacts of operational interventions, and CP offers mathematically assured, distribution-free uncertainty intervals. Our empirical assessment using real-world operational data reveals a distinct differentiation in model performance: while GNN-enhanced hybrids excel in aggregate prediction tasks (CV R² ≈ 0.87), the proposed CML-CP framework realizes a transformative, order-of-magnitude decrease in causal effect estimation error (CV MAE: 124,758.04). Thus, the primary contribution of this research is not merely a singular “best” model, but rather a methodological roadmap that facilitates a paradigm shift from reactive data modeling to a proactive, strategic decision-support tool. This framework equips decision-makers to perform reliable what-if scenario analyses, supported by robust causal insights and valid uncertainty guarantees, leading to more resilient and efficient railway operations.