Answers are grounded in this paper and cite their sources.
Computational study demonstrates enhanced resource utilization and reduced waiting times in simulated tourism environments, suggesting scalable real-time spatiotemporal decision-making.
Efficient service resource allocation in smart tourism environments requires accurate perception of dynamic spatiotemporal information and adaptive decision-making under uncertain conditions. This study proposes a deep reinforcement learning framework that formulates scenic resource scheduling as a partially observable Markov decision process (POMDP) and implements optimization using a customized Rainbow Deep Q-Network (Rainbow DQN). The proposed architecture integrates Distributional Reinforcement Learning, Dueling Network, Prioritized Experience Replay, Multi-step Learning, Double DQN, and Noisy Networks with a hybrid CNN–LSTM feature extraction module to jointly model spatial dependencies and temporal evolution from heterogeneous environmental observations. A multiobjective reward function is designed to simultaneously minimize visitor waiting time, improve service resource utilization, and reduce operational costs while penalizing service violations. Extensive simulations under multiple visitor -flow and environmental scenarios demonstrate that the proposed method consistently outperforms representative reinforcement learning baselines, achieving visitor waiting times of 3.9–8.2 min and resource utilization rates of 76.9%– 88.3% with strong robustness under adverse conditions. Beyond intelligent tourism management, the proposed spatiotemporal decision framework provides methodological insights for adaptive resource scheduling in distributed sensing systems, wireless communication infrastructures, and electromagnetic information acquisition networks, where reliable real-time perception and dynamic optimization are essential for efficient system operation.
No takes yet. Share an insight, caveat, or question.
Xiong et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: