With the advancement of agricultural modernization, agricultural machinery is widely used for crop harvesting. Traditionally, agricultural machines must be refueled at gas stations regularly, affecting the harvesting efficiency. A mobile refueling service has emerged in recent years, in which refueling tankers can move to serve the refueling request. However, the current mobile refueling system is still in an on-demand mode, which may not achieve timely response. Therefore, in this paper, we propose a new mobile refueling mode, i.e., predictive mobile refueling. To tackle the challenge of sparse rewards in predictive mobile refueling, we develop a two-stage reinforcement learning-based scheduling strategy MobRef, which decouples the scheduling process into a central request dispatcher and a distributed tanker reposition scheduler, and further introduces a potential energy-based reward shaping function to facilitate the training of the reposition scheduler. Extensive experiments on two real-world datasets demonstrate the effectiveness of MobRef, which outperforms the best baseline by 12.71% on average. We also present a deployed system based on MobRef, which is used internally in China National Petroleum Corporation.

Time synchronization is a critical requirement in sixth-generation (6G) communication systems and Internet of Things (IoT) applications. However, the standard timing services provided by global navigation satellite systems (GNSS) are increasingly unable to meet the growing demands for precision and stability in modern networks. The proposed system leverages a single-difference time transfer (SDPT) method to achieve tens of picoseconds precision, enabling effective clock disciplining for oven-controlled crystal oscillators (OCXO). A clock disciplining within a phase-locked loop (PLL) structure further optimizes the output of time and frequency signals, enhancing OCXO stability. Additionally, a differential precise time synchronization system (DPTSS) by the BeiDou navigation satellite system (BDS)/global positioning system (GPS) is developed to deliver high-precision time reference signals for communication networks. Experimental results demonstrate that the real-time 1 pulse-per-second (1PPS) interval achieved by the BDS/GPS time link is better than 0.1 ns, with long-term frequency stability reaching the$10{^{-}16 }$level. Furthermore, the proposed system achieves an 1PPS signal synchronization precision of 0.73 ns in communication networks, significantly outperforming the other state-of-the-art methods.

Emotion-cause pair extraction (ECPE), which aims at simultaneously extracting emotion-cause pairs that express emotions and their corresponding causes in a document, plays a vital role in understanding natural languages.Considering that a cause usually appears around its corresponding emotion, we construct a pair graph and a Pair Graph Convolutional Network (PairGCN) to model dependency relations among local neighborhood candidate pairs.Moreover, in our proposed graph, there are three types of dependency relations and each type of dependency relations has its own way to propagate contextual information.Experiments on a benchmark Chinese emotion-cause pair extraction corpus demonstrate the effectiveness of the proposed model.

Spatial information technology has been the main strength to improve Xinjiang's environment and agriculture sustainable development. According to the specific characteristics and requirements, and the serious problems in face of Xinjiang, the paper studied the application mode of spatial information technology for precision agriculture. To realize the development strategic, an integrated PAMS (Precision Agriculture Monitoring and Servicing System in Xinjiang) system was brought forward and designed to improve the utilization of spatial information for ecosystem and environment protection, and agricultural production (precision agriculture). Shihezi was selected as the demonstration area. The PAMS system included spatial information acquiring, processing, analyzing, decision making and service platforms (systems). The aerial remote sensing system would be an important platform and tools for high resolution requirements and emergent events especially, and has been placed to the preferential position. The system, proposed to be developed in Shihezi city, could provide a possible approach for resolving the serious problems in face of Xinjiang.

When people make use of the limited, expensive and historical data to build multiple-input and multiple-output nonlinear mathematical model for decision-making, they often face the problems whether or not all of the experimental data can be used directly for modeling, although artificial neural network (ANN) is a good method to describe the non-linear relationship between inputs and outputs. In the paper, decision-making modeling method based on feed forward ANN and data envelopment analysis (DEA) is brought forward. Experimental data were evaluated and projected by DEA, a widely used method to evaluate relative efficiency among decision making units (DMU). Then the experimental data would become more scientific and reasonable, and all of them could be used for decision-making modeling of ANN. Experiments show that the model of ANN, which gained by training these data, is DEA effective. So it is a new method for optimal data utilizing and decision-making modeling. The method is useful to the research, which may only get limited and high cost data after several times or several years of experiments.