Regulators adopt incentive strategies to guide urban travel behavior, but travelers deviate from system-optimal routes due to heterogeneous information access and perceptual biases. Traffic assignment quantifies such regulatory impacts on network flow, yet fixed user equilibrium-based research overlooks dynamic incentive penetration and random perceptual responses to incentives. This study explores incentive mechanisms’ effects on route choice and traffic assignment, developing two extended equilibrium models: a Mixed User Equilibrium (MUE) model integrating reward penetration and dynamically adjusted incentives, and a Stochastic User Equilibrium (SUE) model with an adaptive perception coefficient for characterizing random perceptual biases and learning effects. Tailored Frank-Wolfe algorithms are designed to solve both models. Numerical experiments on Braess and Sioux Falls networks verify that rational incentive levels and penetration optimize flow distribution and reduce total network impedance, and the SUE model outperforms the MUE model in reflecting actual travel behavior under high-demand network scenarios.
Traffic safety risks of low-speed electric vehicles (LSEVs) among older drivers are a growing concern. Addressing limited existing research, this study investigates injury severity factors and unobserved heterogeneity using 1,346 older-driver LSEV crashes in Zibo, China (2022–2023). Following variable selection via Extreme Gradient Boosting (XGBoost) and Shapley Additive Explanations (SHAP), a random parameter logit model with mean and variance heterogeneity was developed. The model demonstrated superior goodness-of-fit compared to conventional models. Results show airbags, cardiovascular diseases, and vehicle type significantly affect injury severity. Licensing status and driver age were identified as significant random parameters, with several variables affecting their means and variances. This study comprehensively identifies risk factors, providing a scientific basis for targeted LSEV safety policies.
For short-term traffic flow prediction to support signal timing, this paper follows a three-stage procedure: sensing, predicting, and evaluating. At the sensing stage, an adaptive Gaussian mixture model with a detection zone and morphological filtering reduces repeated counting. A dual closed-loop system improves robustness under illumination changes, achieving 90.28% counting accuracy on Dataset A from Xiangyang, China. At the prediction stage, eight deep learning models are compared: six baselines (LSTM, CNN, CNN-LSTM, CNN-LSTM-Attention, LSTM-Transformer, CNN-LSTM-Transformer) and two optimization-enhanced variants (THRO-LSTM, CPO-LSTM), plus GA, PSO, and Bayesian optimization. The CRITIC-TOPSIS method ranks models by accuracy, efficiency, and robustness. Two datasets are used: Dataset A for fitting and Dataset B for generalization. THRO-LSTM achieves a MAPE of 2.12% and a 20% MSE reduction over baseline LSTM in offline scenarios, while plain LSTM performs well in real-time settings. Sensitivity analysis confirms robustness. This study offers a reference for model selection in traffic prediction.
Metro systems move large numbers of commuters, but limited capacity can leave passengers stranded during peak periods, increasing platform density and safety risks. This study proposes a passenger inflow control strategy using a PID controller driven by two parameters, phi and |lambda|, derived from complex eigenvalue analysis of the elliptic bivariate relationship between pedestrian accumulation and time spent in the paid area. Here, phi identifies whether congestion is worsening or easing, while |lambda| represents its severity. The controller estimates the net inflow and supports the allocation of entry and exit turnstiles in AFC systems. The strategy was implemented and evaluated in VISWALK using a calibrated real-world metro station. Compared with no control, it reduced average time spent by 10.2% and stranded passengers by 18.7%, and it also outperformed a threshold-based benchmark in average operational performance. Robustness tests under step disturbances and gain perturbations confirmed stable performance under non-recurrent demand conditions.
