This study focuses on optimizing cache strategies in backend systems and proposes an adaptive cache eviction mechanism based on deep reinforcement learning. Traditional algorithms often struggle with complex and dynamic access patterns. They lack state awareness and the ability to evolve their policies. To address these limitations, this study models cache management as a Markov Decision Process and introduces Deep Q-Networks (DQN) as the decision-making core. An agent is built to learn the optimal actions under different cache states. The model extracts key features such as cache hit rate, access frequency, and time intervals through state representation. It uses experience replay and a target network to ensure stable training. The method is evaluated in several typical experimental scenarios, including hot and cold data switching, changes in cache capacity, and dynamic request distributions. In all cases, the method shows strong adaptability and stability in metrics such as hit rate, response time, and eviction efficiency. The experimental results show that the method effectively improves cache utilization and reduces system latency. It also demonstrates the ability to continuously learn and adjust policies in complex environments. The study provides a systematic evaluation from three perspectives: method modeling, system implementation, and experimental analysis. It confirms the feasibility and practical value of applying deep reinforcement learning to backend cache management. This work offers technical support for intelligent cache optimization in highly dynamic service environments.

This paper proposes a Retrieval-Augmented Generation (RAG) model that integrates structured and unstructured knowledge. The aim is to enhance the knowledge coverage and generation accuracy of large language models in complex question-answering tasks. The method introduces a dual-channel knowledge retrieval mechanism. One channel targets structured knowledge sources such as knowledge graphs and databases. The other focuses on unstructured textual resources such as documents and paragraphs. A unified knowledge fusion network integrates both types of heterogeneous information into a coherent generation context. The model performs multi-level modeling across key components. These include query representation generation, knowledge retrieval, representation alignment, and fusion expression construction. As a result, the generation stage produces text that is semantically rich and logically consistent. Under low-resource conditions, the method significantly improves the accuracy and linguistic quality of generated outputs. It also shows strong stability and generalization in cross-domain tasks. Systematic experiments were conducted on the proportion of structured knowledge, types of knowledge sources, and fusion strategies. The results demonstrate the effectiveness of the fusion architecture in enhancing knowledge representation in language models. This study provides a methodological foundation and empirical support for building controllable and trustworthy knowledge-enhanced natural language generation systems.

This study focuses on the challenge of predicting network traffic within complex topological environments. It introduces a spatiotemporal modeling approach that integrates Graph Convolutional Networks (GCN) with Gated Recurrent Units (GRU). The GCN component captures spatial dependencies among network nodes, while the GRU component models the temporal evolution of traffic data. This combination allows for precise forecasting of future traffic patterns. The effectiveness of the proposed model is validated through comprehensive experiments on the real-world Abilene network traffic dataset. The model is benchmarked against several popular deep learning methods. Furthermore, a set of ablation experiments is conducted to examine the influence of various components on performance, including changes in the number of graph convolution layers, different temporal modeling strategies, and methods for constructing the adjacency matrix. Results indicate that the proposed approach achieves superior performance across multiple metrics, demonstrating robust stability and strong generalization capabilities in complex network traffic forecasting scenarios.