Although Adaptive Traffic Signal Control (ATSC) can alleviate congestion issues to some extent in traditional signal control systems, it still faces challenges in dealing with complex and dynamic traffic environments, such as difficulties in agent coordination, high computational complexity, and unstable optimization results. To address these challenges, this paper proposes a multi-agent deep reinforcement learning algorithm based on SENet, called SE-A3C. The SE-A3C algorithm enhances the feature extraction capability and adaptability of the neural network by introducing the Squeeze-and-Excitation (SE) module from SENet. This allows the model to focus more precisely on high-information features and capture interdependencies between different channels, thereby improving the model’s discriminative ability and decision-making performance. Additionally, the algorithm incorporates Nash equilibrium concepts to maintain a relative balance among agents during coordinated control, avoiding suboptimal competition between agents and significantly improving system stability and efficiency. Experimental results show that, compared to traditional A3C, DQN, and Ape-X algorithms, the SE-A3C algorithm significantly improves the efficiency of traffic signal control and the overall throughput of traffic flow in complex traffic scenarios.

The rise of “Internet + Recycling” platforms is transforming the domestic waste management landscape, creating dual-channel reverse supply chains where new platforms interact with traditional recyclers. However, these platforms face critical strategic decisions regarding their service portfolios (convenient but costly door-to-door vs. economical fixed-point drop-off) and their relationship with incumbents (cooperation vs. competition). This study aims to determine the optimal pricing, service level, and relationship strategies for an “Internet + Recycling” center to maximize profitability under the influence of consumer channel preferences and government subsidies. We developed four Stackelberg game-theoretic models representing different scenarios of service modes (fixed-point only vs. fixed-point with door-to-door) and relationship structures (cooperation vs. competition). We derived equilibrium solutions for recycling prices, service levels, and profits. Our results reveal that while cooperation generally leads to higher systemic profits, the addition of a door-to-door service significantly alters the strategic landscape. We find that a higher consumer preference for the platform channel allows the center to lower prices while increasing profits, and that government subsidies are the most effective at enhancing service levels in cooperative models. Crucially, intense competition incentivizes recycling centers to reduce rather than increase their service levels to cut costs. This research provides a decision-making framework for recycling enterprises to select optimal service and competitive strategies. It also offers insights for policymakers on how to design subsidies to effectively promote high-convenience recycling services and foster a more efficient circular economy.

In the era of rapid urban digitalization, intelligent analysis systems are pivotal for urban safety management, yet their applicability lacks a robust evaluation framework. This study develops a comprehensive model integrating the Delphi method, Fuzz...

In hazardous material transportation route planning, complex environments, stringent safety requirements, and sparse rewards often cause traditional reinforcement learning methods to converge slowly and yield unstable policies. To address these issues, this paper proposes TAR-RL (Trend-Aware and Attention-Augmented Reward Redistribution Reinforcement Learning), which integrates reward redistribution and adaptive exploration to improve the efficiency and robustness of path planning. Specifically, a dynamic ε-greedy strategy based on Q-value trends is introduced to adaptively adjust the exploration rate. Meanwhile, the RUDDER framework decomposes and redistributes delayed rewards to strengthen learning signals at critical decision points. Furthermore, a multi-head attention mechanism is incorporated to model long-term dependencies between states and actions, improving return prediction accuracy. An intrinsic reward mechanism encourages proactive exploration and autonomous learning. Experiments on two grid environments with different complexity levels demonstrate that TAR-RL achieves faster convergence, higher success rates, and more stable training dynamics than baseline methods. These results indicate that TAR-RL is well-suited for safety-aware path planning in high-risk routing scenarios.