Fengshui is extensively employed in China to determine the best locations for ancient buildings, villages, palaces, and tombs. Exploring the science of Fengshui is vital in improving the quality of urban planning, site selection, and human environment. We take the “Form School” Fengshui as the research object, take the Location Selection of Hakka Villages as an Example, and propose a comprehensive site selection model (AHP-GIS model) by combining expert consultation, hierarchical analysis, spatial superposition analysis and kernel density analysis. According to the geographical single-factor evaluation, AHP-GIS evaluation, and the distribution results of traditional Hakka villages in Ganzhou, China, a comparative analysis is conducted to verify the scientific nature of Fengshui. We also evaluate the Science of Fengshui from the micro perspective by the example of the site selection of Bailu Ancient Village. The results show that the geographic single-factor evaluation results and AHP-GIS comprehensive evaluation results coincide with the spatial distribution of Hakka traditional villages. The Fengshui has played an important guiding role in the site selection of Hakka traditional villages in Ganzhou. Its basic principles contain rich knowledge of geography, ecology, psychology and sociology, which has important reference value for guiding urban planning and construction and improving human settlements.

Particle swarm optimization (PSO) algorithms have been successfully used for various complex optimization problems. However, balancing the diversity and convergence is still a problem that requires continuous research. Therefore, an evolutionary experience-driven particle swarm optimization with dynamic searching (EEDSPSO) is proposed in this paper. For purpose of extracting the effective information during population evolution, an adaptive framework of evolutionary experience is presented. And based on this framework, an experience-based neighborhood topology adjustment (ENT) is used to control the size of the neighborhood range, thereby effectively keeping the diversity of population. Meanwhile, experience-based elite archive mechanism (EEA) adjusts the weights of elite particles in the late evolutionary stage, thus enhancing the convergence of the algorithm. In addition, a Gaussian crisscross learning strategy (GCL) adopts crosslearning method to further balance the diversity and convergence. Finally, extensive experiments use the CEC2013 and CEC2017. The experiment results show that EEDSPSO outperforms current excellent PSO variants.

In this study, ethylenediamine (EDA)-functionalized magnetic microspheres with large mesopores (MMMs) were fabricated and explored for the use as highly efficient adsorbents to remove Congo red (CR). The size of mesoporous channels and loading of EDA functional groups were adjusted to optimize the CR adsorption performance of resulting adsorbents. It was found that higher specific surface area (SBET) was favorable for greater CR adsorption, while increasing EDA loading resulted in CR adsorption capacity with a change of first increasing and then declining. The optimal adsorbent FSN-2 with 1.71 wt% N content possessed an ultrahigh maximum CR adsorption capacity of 1662 mg/g, ascribed to its high SBET of 530.91 m2/g and large average pore size of 8.78 nm. The experimental data were fitted best by using the Langmuir model and pseudo-second-order model, suggesting that the adsorption feature be monolayer and chemosorption. The thermodynamic parameters, i.e.Delta G, Delta H and Delta S, were also estimated, indicating the CR adsorption onto FSN-2 was spontaneous and endothermic in nature. High adsorption capacities were recorded in a wide pH range of 5.0 - 9.0, while the addition of Cl-, SO42and NO3- had negligible effect on its CR removal. Moreover, FSN-2 showed good reusability and high stability in adsorption-desorption cycles, as well as excellent CR removal capacity in simulated actual wastewater. The adsorption mechanisms of CR onto FSN-2 mainly involved electrostatic attraction and hydrogen bonding. This work provides a new insight into the fabrication of highly efficient magnetic adsorbents by optimizing the porous structure and functional group loading for potential practical application in wastewater treatment.

粒子群优化算法因其支配参数少、收敛速度快、易于实现等特点被广泛应用,但是粒子群优化算法存在精度低、容易陷入局部优化的问题.为此提出一种基于双种群交叉学习的粒子群优化算法.在该算法中,整个种群被分为普通子种群和精英子种群.普通子种群采用综合变异机制,该机制通过设置概率参数使普通子种群随机选择朝着优秀粒子的方向或者保持自身方向进行变异,以侧重寻找可能解区域.精英子种群则采用交叉学习机制,将粒子的历史最优和全局最优个体进行交叉生成范例,从而引导粒子对可能解区域进行局部搜索,还提出了一种非线性惯性权重来平衡粒子的全局勘探和局部开发能力.为了验证算法的有效性,在十六个基准问题上进行测试并与其他七种粒子群优化算法变体比较,实验结果表明该算法在求解精度和收敛速度总体排名第一,验证了该算法求解性能优于其他粒子群优化算法变体.

Photocatalysis has been regarded as a promising technology for degrading organic pollutants in wastewater and producing hydrogen. In this paper, TiO2 nanoparticles (NPs) were synthesized to improve the visible light absorption of TiO2, which were further combined with Bi2O3 nanosheets to synthesize a series of 0D/2D TiO2 NPs/Bi2O3 nanosheet heterojunctions. The visible light induced photocatalytic activities of the as-synthesized TiO2/Bi2O3 heterojunctions were studied. The optimized catalyst TB-3 with 15 wt% of Bi2O3/TiO2 exhibited the best photocatalytic degradation of tetracycline hydrochloride (TC). The degradation rate constant k of TC over TB-3 was approximately eight times and 39 times greater than that of P25 and Bi2O3, respectively. Additionally, TB-3 showed the highest amount of hydrogen evolution, while that of Bi2O3 was almost zero. The enhancement of photocatalytic performances was ascribed to the improved visible light absorption and the Z-scheme charge transfer path of the TiO2/Bi2O3 heterojunctions, which enhanced the separation efficiency and reduced recombination of photogenerated charge carries, as evidenced by UV–Visible diffuse reflectance spectroscopy (DRS), photoluminescence spectroscopy (PL), and electrochemistry measurements. The active species trapping experiments and the electron spin resonance (ESR) results revealed that ·O2− was the main active substance in the photocatalytic degradation. The possible degradation pathway and intermediate products of TC have been proposed. This work provides experimental evidence supporting the construction of Z-scheme heterojunctions to achieve excellent visible light induced photocatalytic activity.

Novel rose-like photocatalysts were constructed by 1D/2D La(OH)3/(BiO)2OHCl heterojunctions. The highly efficient photocatalytic degradation was boosted by the combined effects of rich oxygen vacancies and enhanced interfacial charge transfer.

针对传统果树单株提取方法流程繁琐、人力成本较高、耗时长等问题,该文提出在复杂 自然背景下使用深度学习算法YOLOv4实现脐橙树无人机影像端到端的识别方法,利用网络改进、多尺度融合以及损失函数改进等方法提高复杂背景下小目标的识别率.以赣南某地脐橙果树无人机影像作为数据源,应用YOLOv4深度神经网络分别训练基于不同数据集的模型,通过调整模型阈值参数得到最佳模型,并在强背景、弱背景、稀疏和稠密植株测试集上进行测试.结果表明,基于data748数据集训练且模型阈值为0.6时的YOLO-748模型精确率达91.55％,召回率为98.55％,mAP值为93.38％,F1值达0.985,该模型在复杂自然背景下鲁棒性较好.该方法能为现代农业果园管理提供新的可行方案.

Federated learning is a method of distributed machine learning based on privacy protection, which is characterized by a large number of heterogeneous clients, and often these clients have great differences in communication and computation capabilities, such as servers and smartphones. In this work, we propose a new federated learning algorithm named AAFL, which is based on asynchronous communication and can adaptively assign training models based on the capabilities of different clients to solve the communication efficiency and privacy problems caused by a large number of heterogeneous clients. Our proposed solution can assign training models with different computational complexity according to clients with different capabilities, and reduce the problem that some clients with poor capabilities slow down the training speed of the global model by means of synchronous communication. AAFL can allocate sub-models of different sizes and still aggregate a single global model. The experimental results show that assigning models of different scales according to the capabilities of different clients and using asynchronous communication can improve the training speed and accuracy of the global model to a certain extent, and has a good effect on protecting the privacy of user data.

揭示城镇居民生活用水量的变化规律及其影响因素,对于改善城市供水调度和控制居民用水量的增长具有重要意义.为克服BP神经网络学习收敛速度慢、易陷入局部极小值、网络结构难以确定的缺点,提出将遗传算法(GA)与BP神经网络组合形成GA-BP混合训练网络,应用于居民生活用水量预测当中,预测结果显示,该方法与单一 BP神经网络相比,具有更好的预测精度和适用性.其次,通过修正后的模型分析水价、人均可支配收入及人均住房面积三个用水需求影响因素,估计其价格弹性、收入弹性和人均住房需求弹性.模型估计结果表明,水价的上调,会抑制居民用水量的增长;居民用水量明显随人均可支配收入的增加而提升;人均住房面积与居民用水量之间的关系未得到较好体现.

Ciphertext-Policy Attribute-Based Encryption technology binds ciphertext to attribute sets, which can realize fine-grained access control centered on data users. At present, research on attribute encryption mostly focuses on ordinary data, and there is no research on multimedia such as images. Therefore, an image-oriented Ciphertext-Policy Attribute-Based Encryption (CP-ABE) scheme is proposed. Firstly, the image is processed by wavelet decomposition, and four groups of wavelet coefficient matrices are obtained, which will be scrambling encrypted respectively, and then the encrypted image is obtained through wavelet reconstruction. Secondly, the key used for scrambling encryption is encrypted by Ciphertext-Policy Attribute-Based Encryption respectively. Finally, build a demo system on this basis. The experimental results show that there are differences in the quality of the images that can be accessed depending on different user attributes, which realize the fine-grained control of the users' decryption abilities.

导师团队制研究生培养模式有助于避免当前单一导师制研究生培养模式的局限与弊端.本文以化学工程专业研究生为例,分别从高水平研究生导师团队的组建、导师团队的队伍培养和建设、导师团队管理制度的健全、导师团队制下研究生管理体系的构建等方面探讨了基于化学化工学院研究所的导师团队制研究生培养模式的改革和实施对策.初步实践证明,导师团队制研究生培养模式改革,有利于提高化学工程专业研究生的科研创新能力.

In December 2019, corona virus disease 2019 (COVID-19) has broken out in China. Understanding the distribution of disease at the national level contributes to the formulation of public health policies. There are several studies that investigating the influencing factors on distribution of COVID-19 in China. However, more influencing factors need to be considered to improve our understanding about the current epidemic. Moreover, in the absence of effective medicine or vaccine, the Chinese government introduced a series of non-pharmaceutical interventions (NPIs). However, assessing and predicting the effectiveness of these interventions requires further study. In this paper, we used statistical techniques, correlation analysis and GIS mapping expression method to analyze the spatial and temporal distribution characteristics and the influencing factors of the COVID-19 in mainland China. The results showed that the spread of outbreaks in China’s non-Hubei provinces can be divided into five stages. Stage I is the initial phase of the COVID-19 outbreak; in stage II the new peak of the epidemic was observed; in stage III the outbreak was contained and new cases decreased; there was a rebound in stage IV, and stage V led to level off. Moreover, the cumulative confirmed cases were mainly concentrated in the southeastern part of China, and the epidemic in the cities with large population flows from Wuhan was more serious. In addition, statistically significant correlations were found between the prevalence of the epidemic and the temperature, rainfall and relative humidity. To evaluate the NPIs, we simulated the prevalence of the COVID-19 based on an improved SIR model and under different prevention intensity. It was found that our simulation results were compatible with the observed values and the parameter of the time function in the improved SIR model for China is a = − 0.0058. The findings and methods of this study can be effective for predicting and managing the epidemics and can be used as an aid for decision makers to control the current and future epidemics.