Extracting Takagi-Sugeno Fuzzy Rules with Interpretable Submodels via Regularization of Linguistic Modifiers

In this paper, a method for constructing Takagi-Sugeno (TS) fuzzy system from data is proposed with the objective of preserving TS submodel comprehensibility, in which linguistic modifiers are suggested to characterize the fuzzy sets. A good property held by the proposed linguistic modifiers is that they can broaden the cores of fuzzy sets while contracting the overlaps of adjoining membership functions (MFs) during identification of fuzzy systems from data. As a result, the TS submodels identified tend to dominate the system behaviors by automatically matching the global model (GM) in corresponding subareas, which leads to good TS model interpretability while producing distinguishable input space partitioning. However, the GM accuracy and model interpretability are two conflicting modeling objectives, improving interpretability of fuzzy models generally degrades the GM performance of fuzzy models, and vice versa. Hence, one challenging problem is how to construct a TS fuzzy model with not only good global performance but also good submodel interpretability. In order to achieve a good tradeoff between GM performance and submodel interpretability, a regularization learning algorithm is presented in which the GM objective function is combined with a local model objective function defined in terms of an extended index of fuzziness of identified MFs. Moreover, a parsimonious rule base is obtained by adopting a QR decomposition method to select the important fuzzy rules and reduce the redundant ones. Experimental studies have shown that the TS models identified by the suggested method possess good submodel interpretability and satisfactory GM performance with parsimonious rule bases.     

基于规则引擎的智能决策支持系统开发

提出了解决基于规则专家系统的规则冲突的方法。构建了使用Eclipse和Drools的集成开发环境。     

Detection of urinary biomarkers for early diagnosis of acute renal allograft rejection by proteomic analysis

Acute allograft rejection has been recognized as a major impediment to improved success in renal transplantation. Timely detection and control of rejection are very important for the improvement in long-term renal allograft survival. Thus, biomarkers for early diagnosis of acute rejection are required urgently to clinical medication. This study seeks to search for such biomarker candidates by comparing patients' pre-treatment urinary protein profiling with their post-treatment urinary protein profiling. A total of 15 significantly and consistently down-regulated protein candidates were identified. Among them, alpha-1-antichymotrypsin precursor (AACT), tumor rejection antigen gp96 (GP96) and Zn-Alpha-2-Glycoprotein (ZAG) were selected for further analysis. The results indicated that Western Blot assay of AACT, GP96 and ZAG had advanced the diagnosis time of acute renal rejection by 3 days, compared with current standard clinical observation and laboratory examination. Furthermore, the double-blind detection revealed that the accuracy, sensitivity and specificity of the diagnosis of acute renal rejection of AACT, GP96 and ZAG were 66.67%/100%/60%, 83.33%/100%/80% and 66.67%/100%/60%, respectively, and 100%/100%/100% in combination. In conclusion, urinary protein AACT, GP96 and ZAG could be a set of potential biomarkers for early non-invasive diagnosis of the acute rejection after renal transplantation.     

基于网格的一种新的动态演化算法

近年来,越来越多的演化计算研究者对动态优化问题产生了很大的兴趣,并产生了很多解决动态优化问题的方法。提出一种新的动态演化算法,与传统的演化算法有所不同,它是建立在划分网格基础上的,故而称它为网格优化算法。通过测试典型的动态优化问题,并与经典的SOS算法进行比较,证明了算法的有效性。     

基于概率A*的智能车路径规划算法

为保证智能车辆能够按照规划的路径安全行驶,且满足车辆动力学特性,针对混合A*算法中没有引导性策略所产生较多不必要的搜索问题,提出概率A*算法先得到搜索粗路径,提高在随后搜索过程中搜索效率。然后利用概率A*算法得到的路径点引导节点走向,避免节点向障碍物等搜寻。最后对节点的代价函数进行优化。仿真实验结果表明,与混合A*算法相比,本文所提出的算法平均减少搜索时间10.8%,且得到的路径相对规整平滑。该算法可以在较短时间内为智能车辆规划一条安全可行平滑的路径。     

Simultaneous CO2 and H2O Activation via Integrated Cu Single Atom and N Vacancy Dual-Site for Enhanced CO Photo-Production

Photocatalytic conversion of CO₂ into fuels using pure water as the proton source is of immense potential in simultaneously addressing the climate-change crisis and realizing a carbon-neutral economy. Single-atom photocatalysts with tunable local atomic configurations and unique electronic properties have exhibited outstanding catalytic performance in the past decade. However, given their single-site features they are usually only amenable to activations involving single molecules. For CO₂ photoreduction entailing complex activation and dissociation process, designing multiple active sites on a photocatalyst for both CO₂ reduction and H₂O dissociation simultaneously is still a daunting challenge. Herein, it is precisely construct Cu single-atom centers and two-coordinated N vacancies as dual active sites on CN (Cu₁/N_2CV-CN). Experimental and theoretical results show that Cu single-atom centers promote CO₂ chemisorption and activation via accumulating photogenerated electrons, and the N_2CV sites enhance the dissociation of H₂O, thereby facilitating the conversion from COO* to COOH*. Benefiting from the dual-functional sites, the Cu₁/N_2CV-CN exhibits a high selectivity (98.50%) and decent CO production rate of 11.12 µmol g⁻¹ h⁻¹. An ingenious atomic-level design provides a platform for precisely integrating the modified catalyst with the deterministic identification of the electronic property during CO₂ photoreduction process.     

Green Synthesis of Hierarchically Porous Carbon Nanotubes as Advanced Materials for High‐Efficient Energy Storage

Hierarchically porous carbon nanomaterials with well‐defined architecture can afford a promising platform for effectively addressing energy and environmental concerns. Herein, a totally green and straightforward synthesis strategy for the fabrication of hierarchically porous carbon nanotubes (HPCNTs) by a simple carbonization treatment without any assistance of soft/hard templates and activation procedures is demonstrated. A high specific surface area of 1419 m² g^?1 and hierarchical micro‐/meso‐/macroporosity can be achieved for the HPCNTs. The unique porous architecture enables the HPCNTs serving as excellent electrode/host materials for high‐performance supercapacitors and Li–sulfur batteries. The design strategy may pave a new avenue for the rational synthesis of hierarchically porous carbon nanostructures for high‐efficient energy storage applications.     

Ultraviolet Photoluminescence of Carbon Nanospheres and its Surface Plasmon‐Induced Enhancement

Ultraviolet (UV) light can be used in versatile applications ranging from photoelectronic devices to biomedical imaging. In the development of new UV light sources, in this study, stable UV emission at ≈350 nm is unprecedentedly obtained from carbon nanospheres (CNSs). The origin of the UV fluorescence is comprehensively investigated via various characterization methods, including Raman and Fourier transform infrared analyses, with comparison to the visible emission of carbon nanodots. Based on the density functional calculations, the UV fluorescence is assigned to the carbon nanostructures bonded to bridging O atoms and dangling –OH groups. Moreover, a twofold enhancement in the UV emission is acquired for Au‐carbon core‐shell nanospheres (Au‐CNSs). This remarkable modification of the UV emission is primarily ascribed to charge transfer between the CNSs and the Au surface.     

宽翼T形梁桥动力学理论与特性分析

考虑了剪滞翘曲应力自平衡条件、剪切变形和剪力滞后效应等因素的影响,本文提出了一种对宽翼薄壁T形梁动力学特性的分析方法.分析中为了准确反应T形梁翼板的动位移变化,三个广义动位移被引入,且以能量变分原理为基础建立了T形梁动力反应的控制微分方程和自然边界条件,据此对T形梁的动力反应特性进行了分析,揭示了T形梁桥动力反应的规律.算例中,对比了考虑和不考虑剪滞翘曲应力自平衡条件对T形梁动力反应的影响,结果显示考虑剪滞翘曲应力自平衡条件的计算方法与有限元数值解吻合更好.