基于支持向量机的织物悬垂性能评估分析

 针对织物悬垂性评估中存在非线性建模困难及评估精度不高等问题，结合织物悬垂性参数数据的特点，基于Mercer核函数的性质，构建Mercer核函数的织物悬垂性评估模型。通过实验，认为其评估精度有了相应的提高，说明基于Mercer核函数的织物悬垂性评估模型是可行的。同时，通过改变参数值进行实验，结果表明这些参数的取值不同对评估结果有影响，且不同参数对评估结果影响的程度也不同，这对基于支持向量机的织物悬垂性评估过程有重要的指导作用。  

Optimization of the enzymatic pretreatment in oat bran protein extraction by particle swarm optimization algorithms for response surface modeling

Enzymatic pretreatments of oat bran, using viscozyme L to enhance protein extraction, were investigated. To obtain the best possible combination of enzymatic pretreatment variables including viscozyme L concentration, incubation time, pH and incubation temperature for maximum protein extraction, modern optimization techniques were employed. The optimization procedure included two stages: creating a response surface model through regression on experimental data and then optimizing the model by Particle Swarm Optimization (PSO) algorithms. The optimum conditions of enzymatic pretreatment were: viscozyme L concentration, 30 FBG/10 g of oat bran; incubation time, 2.8 h; pH 4.6; and incubation temperature, 44 °C. Under such conditions, the predicted maximum protein extracted was 55.69%, which coincided with the experimental values. These results indicated that enzymatic pretreatment can improve protein extraction significantly and PSO algorithms were efficient approaches for optimizing the conditions of enzymatic pretreatment of oat bran.  

基于区块链的食品安全溯源体系设计

近年来，食品安全案件屡次发生，亟需采用安全、可信、透明的食品安全溯源系统来加强对食品产业链的监管效率，提高食品安全水平，保障国民饮食健康；而目前的食品溯源系统主要面临着信息采集不标准、数据存储不安全、中心系统易受攻击和企业间信息交换过程隐私不能保障等问题。区块链技术具有分布式容错、不可篡改与隐私保护的特点，可以针对性地解决目前食品溯源系统中面临的问题。基于区块链的食品安全溯源系统通过技术架构的分析，提出将区块链恰当植入食品溯源体系的方案，即区块链技术应用于系统的数据库层与通信层，并对采用方案后的食品溯源体系的运行机制进行分析，结合具体应用场景与实际案例来论证设计方案的有效性。  

Metabolic engineering--integrating methodologies of molecular breeding and bioprocess systems engineering

Metabolic engineering is an integrating methodology of analysis and synthesis for the improvement of flux distribution of metabolic pathways in complicated bioprocesses, which are highly multi-hierarchical systems to extend from macroscopic to microscopic levels. Recent progress in metabolic engineering methodologies to improve metabolic pathways in microorganisms was reviewed with many studies in this paper. Metabolic flux distribution was analyzed under different environmental conditions, using a metabolic reaction model. The physiological states of microorganisms were understood by interpreting metabolic flux analysis (MFA). This analysis was also used for development of process operation and control strategy. Cell capability to form a targeted product was analyzed with a metabolic reaction model and linear programming (LP). The use of a 13C-enriched carbon source and nuclear magnetic resonance (NMR) and gas chromatography-mass spectrometry (GCMS) analyses of intracellular and extracellular metabolites enabled determination of a metabolic flux distribution more accurately than the flux distribution determined only by the metabolic reaction model, which involves not only metabolite balances but also energy and redox balances. The comparison of metabolic flux distributions between before and after genetic modification of cells yielded information on the mechanism of regulation of metabolic flux in microorganisms. Finally, integration of bioinformatics and metabolic engineering is discussed, and cyclic modification of the complex bionetwork and process development were emphasized.  

基于多项式回归模型的液晶显示器特性化

为了实现LCD液晶显示器精确的色度特性化,用多项式回归算法建立了液晶显示器RGB空间到CIEXYZ空间的转换模型,详细研究了多项式回归模型不同训练样本、不同项数以及阶数对液晶显示器特性化的影响。实验结果表明,项数小于9时,模型预测的精度较低,项数大于9时模型精度逐渐提高,用三阶三元十四项多项式可以得到平均误差为0.6623左右的高精度预测模型。  

模切工艺优化系统及实现策略研究

随着全球包装印刷行业的迅速发展和社会对印刷精品需求的大幅增长,模切压痕工艺的重要性受到越来越多企业的关注.本研究在对国内外模切压痕工艺研究现状进行分析的基础上,从模切设备的性能和模切组件的选用角度分析了国内外模切工艺的异同,挖掘国内模切工艺存在的不足,提出对模切工艺系统进行优化.利用数据库技术、建模仿真技术和优化设计方法,构建模切工艺虚拟仿真系统,提出了系统开发流程和实现策略,分析了实现虚拟仿真系统的关键技术及存在的相关问题,并对模切工艺的发展趋势进行了展望.  

计算机视觉技术在海产品质量评估中的应用

传统的海产品分类及质量检验主要是采用人工方法,靠肉眼进行判断,需耗费大量的人力,且劳动强度大、效率低。采用计算机视觉技术对海产品进行自动分类与质量评估可有效克服上述缺点。通过提取海产品的尺寸、形状、颜色、纹理等特征,结合预测模型,采用数字图像处理的方法可实现海产品的无接触、无损伤处理。文章分析基于计算机视觉技术的海产品分类及质量评估系统的构成,并通过实例分析阐述其具体实现方法,论证该方法的有效性。  

新疆特色干酪中乳酸菌的分离鉴定

从新疆不同牧区采集工艺不同的干酪制品,对其中的乳酸菌进行分离纯化、生理生化性质试验和16S rRNA分析.结果表明,分离、纯化出的104株乳酸菌种,有82株为乳杆菌属(Lactobacillus),12株为肠球菌属(Enterococcus),10株为魏斯氏菌属(Weissella).利用16S rRNA序列同源分析和系统发育树分析对具有不同生理生化特性的代表菌株进行了分子鉴定,鉴定结果为TNM-2与干酪乳杆菌(Lactobacillus casei)、Y5-4与食窦魏斯氏菌(Weissella cibaria)、NS2-2与植物乳杆菌(Lactobacillus plantarum)的同源性达到100％,NM-2与瑞士乳杆菌(Lactobacillus helveticus)、Y1-1与马乳酒乳杆菌(Lactobacillus kefiranofaciens)、WG与耐久肠球菌(Enterococcus durans)的同源性达到99％.  

制浆造纸工业过程仿真平台的研究

采用序贯模块法和面向对象程序设计的方法对制浆造纸工业过程仿真平台的总体结构进行设计,开发了与浆水衡算和能量衡算有关的功能模块,并通过实例验证了仿真平台的有效性。  

基于广播加密的对称叛逆者追踪方案分析

对基于广播加密的对称叛逆者追踪方案进行了讨论与分析。结果表明:当用户数n不是太大时,对称方案具有高效、简单、安全性高的特性。在对称叛逆者追踪方案中,数字信息的解密密钥并非由用户单方面所知,而是由用户和数据发行商双方所共知。因此,数据发行商完全可以恶意使用此私有密钥,来诬陷诚实的授权用户。所以对称叛逆者方案不具备防诬陷性。针对对称方案的缺陷,研究出非对称方案。在非对称方案中,授权用户的解密密钥仅其本人知道,数据供应商一无所知,从而弥补了对称方案的可否认、可诬陷性。通过对4种具有代表性的对称叛逆者追踪方案的分析,在安全性的基础上,对其进行改进,可进一步研究出具有更多特性的叛逆者追踪方案。