The question concerning how digital consumption demand has been adapted and how matching business models have been built has become an important practical problem in the digital development of the retail industry. Considering the effects of COVID-19, whether new retail enterprises can maintain adequate competitiveness and risk resilience in the post-pandemic era deserves in-depth study. In comparing the development of traditional retail and new retail enterprises, we extracted and evaluated key factors of enterprise operating efficiency. Then, we measured the transformation efficiency of 65 enterprises in China listed in 2016 and 2020 by establishing a DEA model and the Malmquist index method. Finally, based on an empirical analysis demonstrating the necessity of traditional retail transformation, we analyzed retail enterprises’ efficiency and dynamic efficiency changes. The results show that the operating efficiency of enterprises using the new retail model was higher than those using the traditional retail model. The technical efficiency and total factor productivity were significantly improved after the new retail model was applied. Both technological progress and improved technological efficiency contributed to the improvement in total factor productivity during the COVID-19 pandemic.

As one of the most important grain protection policies in China, the minimum purchase price policy prevents the fluctuation of grain output and protects the interests of farmers by regulating the prices of major grain varieties. For developing countries with a shortage of agricultural resources, represented by China, an in-depth study on the implementation effect and public satisfaction of this policy is of great significance for promoting the sustainable development of the grain industry. Based on the interest demands of the government, farmers, grain enterprises and consumers, this paper constructs a policy satisfaction evaluation model based on the Analytic Hierarchy Process and Fuzzy Comprehensive Evaluation. The research shows that the implementation effect of this policy has promoted the sustainable development of China’s grain in four aspects: improving farmers’ enthusiasm for planting, optimizing the structure of supply and demand, reducing the adverse impact of disasters, and ensuring the steady increase of output. However, due to the differences in natural resources and folk customs, the implementation effect of this policy varies in different regions.

Due to limited communication, computing resources, and unstable environments, traditional cold chain traceability systems are difficult to apply directly to marine cold chain traceability scenarios. Motivated by these challenges, we construct an improved blockchain-based cold chain traceability system for marine fishery vessels. Firstly, an Internet of Vessels system based on the Iridium Satellites (IoV-IMS) is proposed for marine cold chain monitoring. Aiming at the problems of low throughput, long transaction latency, and high communication overhead in traditional cold chain traceability systems, based on the Practical Byzantine Fault Tolerance (PBFT) consensus algorithm, a Node-grouped and Reputation-evaluated PBFT (NR-PBFT) is proposed to improve the reliability and robustness of blockchain system. In NR-PBFT, an improved node grouping scheme is designed, which introduces a consistent hashing algorithm to divide nodes into consensus and candidate sets, reducing the number of nodes participating in the consensus process, to lower communication overhead and transaction latency. Then, a reputation evaluation model is proposed to improve the node selection mechanism of NR-PBFT. It enhances the enthusiasm of nodes to participate in consensus, which considers the distance between fishery vessels, data size, and refrigeration temperature factors of nodes to increase throughput. Finally, we carried out experiments on marine fishery vessels, and the effectiveness of the cold chain traceability system and NR-PBFT were verified. Compared with PBFT, the transaction latency of NR-PBFT shortened by 81.92%, the throughput increased by 84.21%, and the communication overhead decreased by 89.4%.

Accurate building energy consumption prediction is essential for efficient energy management and energy optimization. This study utilizes bidirectional long short-term memory (BiLSTM) to automatically extract deep time series features. The nonlinear fitting and high-precision prediction capabilities of Random Forest (RF) and XGBoost models are then utilized to develop a BiLSTM-RF-XGBoost stacked hybrid model. To enhance model generalization and reduce overfitting, a Bayesian algorithm with an early stopping mechanism is utilized to fine-tune hyperparameters, and strict K-fold time series cross-validation (TSCV) is implemented for performance evaluation. The hybrid model achieves a high TSCV average R² value of 0.989 during cross-validation. When evaluated on an independent test set, it yields a mean square error (MSE) of 0.00003, a root mean square error (RMSE) of 0.00548, a mean absolute error (MAE) of 0.00130, and a mean absolute percentage error (MAPE) of 0.26%. These values are significantly lower than those of comparison models, indicating a significant improvement in predictive performance. The study offers insights into the internal decision-making of the model through SHAP (SHapley Additive exPlanations) feature significance analysis, revealing the key roles of temperature and power lag features, and validating that the stacked model effectively utilizes the outputs of base models as meta-features. This study makes contributions by proposing a novel hybrid model trained with Bayesian optimization, analyzing the influence of various feature factors, and providing innovative technological solutions for building energy consumption prediction. It also provides theoretical value and guidance for low-carbon building energy management and application.