A Hybrid Deep Learning Framework with Q-Table Optimization for Well Log Reconstruction

The reconstruction of acoustic (AC) logging curves is of great significance for reservoir evaluation, lithology identification, and velocity modeling, particularly in the presence of missing or degraded logging data. However, conventional reconstruction methods and existing deep learning models often suffer from limited feature representation capability and rely heavily on manual hyperparameter tuning, leading to suboptimal performance. To address these challenges, this study proposes a reinforcement learning-based optimization framework for AC logging curve reconstruction. Specifically, a hybrid deep learning architecture integrating convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM), and an attention mechanism is developed to effectively capture local spatial features, long-range temporal dependencies, and key feature contributions from multi-logging data. Furthermore, a Q-learning-based optimization strategy is introduced to adaptively tune model hyperparameters by formulating the optimization process as a Markov Decision Process (MDP), enabling dynamic and data-driven parameter adjustment. To validate the effectiveness of the proposed method, comparative experiments are conducted using several baseline and optimized models, including CNN–BiLSTM, CNN–BiLSTM–Attention, particle swarm optimization (PSO)-optimized CNN–BiLSTM–Attention, and genetic algorithm (GA)-optimized CNN–BiLSTM–Attention. The results demonstrate that the proposed approach achieves superior reconstruction accuracy for AC curves, with improved convergence efficiency and model stability. In addition, it exhibits stronger robustness and generalization capability under limited data conditions, effectively mitigating the risk of overfitting and local optima. This study provides a novel reinforcement learning-driven solution for AC logging curve reconstruction and offers practical value for intelligent reservoir characterization in complex geological environments.