Urban traffic optimization is challenging due to its dynamic, nonlinear nature, limiting the effectiveness of traditional metaheuristics like WOA and DCWOA. This paper introduces TITAN-WOA, which combines reinforcement learning, topological data analysis, and fractional-order Lévy flight for rapid real-time rerouting, and ATLAS-WOA, which employs Riemannian manifold learning, Mean-Field Games, and CMA-ES for scalable fleet coordination. Both frameworks incorporate a 14-operator enhancement pipeline and outperform existing methods on CEC-2017 benchmarks, with ATLAS-WOA delivering the highest optimization accuracy and TITAN-WOA providing significantly lower computational cost, offering an effective trade-off between performance and efficiency.

This paper presents a hybrid EEG-based stress classification framework using Topological Data Analysis (TDA) on the SAM40 dataset. EEG signals are transformed into high-dimensional point clouds via Takens’ embedding, followed by persistent homology to extract topological features, which are combined with conventional EEG features. The proposed approach achieves peak accuracies of 95.45%, 94.52%, and 94.40% across three stress-inducing tasks, demonstrating the effectiveness of TDA for robust stress detection.

This paper proposes P-COB-SSA, a hybrid trajectory optimization framework for autonomous navigation in obstacle-rich environments. The method integrates Logistic Chaotic Mapping for diverse initialization, Artificial Potential Fields for physics-guided collision-free navigation, and Multi-Strategy Opposition-Based Learning to escape local optima. Evaluations on challenging benchmark landscapes demonstrate superior convergence, robust obstacle avoidance, and near-optimal trajectory efficiency (path tortuosity ≈ 1.06), highlighting its effectiveness for autonomous path planning.