基于量子神经网络的马铃薯早疫病诊断模型

针对马铃薯早疫病智能诊断,将量子计算的态叠加方法和神经网络计算的自适应性结合,提出了将量子神经网络作为马铃薯早疫病诊断模型.该模型隐含层采用多个量子能级的激励函数叠加的量子神经元,有效地解决了病害诊断中模糊决策,在给出的学习算法的训练过程中自适应地确定样本特征数据中的不确定性.此算法能够较好地避免传统神经网络在训练过程中易出现局部极小值的弊端,提高了网络学习速度.仿真结果表明:量子神经网络在马铃薯早疫病诊断中,诊断正确率达到96.5％.     

基于动态性能的车架结构模型分析研究

客车车架的动态特性是影响整车振动舒适性的关键因素。为探讨纵横梁结构对某轻型客车车架动态特性的影响,建立车架的研究性模型,对研究性模型进行多种方案的计算分析,综合各方案的计算结果,确定影响车架动态特性最大的结构特征,以此为基础对原车架进行改进,提高了车架的模态频率。计算结果表明,该研究方法简便有效,可推广应用于对其它结构的研究。     

Google Maps在农业WebGIS中的应用探索

针对目前国内农业WebGIS存在的信息单一、功能单调等问题,提出了将标准的地理信息系统开发平台和Google Maps同时应用于农业WebGIS的方法。利用Google Maps提供的开放地图接口自定义GoogleMap类可以为农业WebGIS系统提供详细的地理位置信息的显示、标注以及路线查询等功能。最后通过SuperMap.IS.NET开发平台与Google Maps实现对小汤山农业基地的信息管理,证明该方法实用可行。     

Soil compaction and forest floor removal reduced microbial biomass and enzyme activities in a boreal aspen forest soil

Soil enzymes are linked to microbial functions and nutrient cycling in forest ecosystems and are considered sensitive to soil disturbances. We investigated the effects of severe soil compaction and whole-tree harvesting plus forest floor removal (referred to as FFR below, compared with stem-only harvesting) on available N, microbial biomass C (MBC), microbial biomass N (MBN), and microbial biomass P (MBP), and dehydrogenase, protease, and phosphatase activities in the forest floor and 0–10 cm mineral soil in a boreal aspen (Populus tremuloides Michx.) forest soil near Dawson Creek, British Columbia, Canada. In the forest floor, no soil compaction effects were observed for any of the soil microbial or enzyme activity parameters measured. In the mineral soil, compaction reduced available N, MBP, and acid phosphatase by 53, 47, and 48%, respectively, when forest floor was intact, and protease and alkaline phosphatase activities by 28 and 27%, respectively, regardless of FFR. Forest floor removal reduced available P, MBC, MBN, and protease and alkaline phosphatase activities by 38, 46, 49, 25, and 45%, respectively, regardless of soil compaction, and available N, MBP, and acid phosphatase activity by 52, 50, and 39%, respectively, in the noncompacted soil. Neither soil compaction nor FFR affected dehydrogenase activities. Reductions in microbial biomass and protease and phosphatase activities after compaction and FFR likely led to the reduced N and P availabilities in the soil. Our results indicate that microbial biomass and enzyme activities were sensitive to soil compaction and FFR and that such disturbances had negative consequences for forest soil N and P cycling and fertility.     

Influence of non-cellulose structural carbohydrate composition on plant material decomposition in soil

The C mineralisation pattern during the early stage of decomposition of plant materials is largely determined by their content of different carbohydrates. This study investigated whether detailed plant analysis could provide a better prediction of C mineralisation during decomposition than proximate analysis [neutral detergent solution (NDF)/acid detergent solution (ADF)]. The detailed analysis included sugars, fructans, starch, pectin, cellulose, lignin and organic N. To determine whether differences in decomposition rate were related to differences in hemicellulose composition, the analysis particularly emphasised the concentrations of arabinose and xylose in hemicelluloses. Carbon dioxide evolution was monitored hourly in soil amended with ten different plant materials. Principal component and regression analysis showed that C mineralisation during day 1 was closely related to free sugars, fructans and soluble organic N components (R ² = 0.83). The sum of non-cellulose structural carbohydrates (intermediate NDF/ADF fraction) was not related to C mineralisation between days 1 and 9. In contrast, a model including starch and protein in addition to the non-cellulose structural carbohydrates, with the hemicelluloses replaced by arabinose and xylose, showed a strong relationship with evolved CO₂ (R ² = 0.87). Carbon mineralisation between days 9 and 34 was better explained by xylan, cellulose and lignin (R ² = 0.72) than by lignocellulose in the ADF fraction. Our results indicated that proximate analyses were not sufficient to explain differences in decomposition. To predict C mineralisation from the range of plant materials studied, we propose a minimum set of analyses comprising total N, free sugars, starch, arabinose, xylan, cellulose and lignin.     

甘蔗施肥技术系统优化研究

本文参照系统工程中的最佳模拟配合法，对甘蔗施肥技术的系统优化研究。通过计算机模拟优化，得出若干组包括高，中，低产量，净产值和甘蔗蔗糖分等综合指标要求的相应配套施肥方案，并对有关问题进行了讨论，以期为甘蔗科学施肥提供新的方法及其理论依据。     

杜仲叶采收时间与采叶量的研究

通过对杜仲叶不同采收方式的研究,结果认为,不同采叶方式对杜仲树体及叶生长均有不同程度的影响:采叶量愈大、采叶次数愈多,树木高、径生长量愈小,新生叶叶面积也愈小.但是,50采叶量组和70采叶量组、1次采叶组和2次采叶组差异不显著.所以,从采叶后对杜仲高、径生长量以及叶片生长的影响来看,一次采叶量以50左右为宜,一年内采叶次数不宜超过2次.     

水稻旱育抛秧壮根立苗技术途径的探讨

针对旱育秧苗在抛植情况下，苗体轻、根盘小且根部难以带上这一情况，应用水稻秧苗健壮素和高吸水材料，不仅促进了旱育秧苗根系的发育，壮大了根群，提高了单株带蘖数，而且使秧苗根部带上量明显增加，高吸水材料包衣还使秧苗根部形成“蓄水球囊”，明显提高了旱青抛秧的立苗速度和抗植伤能力。     

几种土壤铁锰结核对Cr(Ⅲ)的氧化动力学特性

以几种土壤铁锰结核为试验材料 ,采用流动法 ,用权函数方程、抛物线方程、Elvoich方程及一级方程模型对锰矿物氧化Cr(Ⅲ )的动力学过程进行了研究 ,结果表明上述方程拟合的相关系数均达到显著水平 (P <0 .0 5 ) ,并且以权函数方程和抛物线方程拟合所得的标准差较小 ,由该氧化反应先吸附再氧化的反应机理、拟合方程的特性 ,表明氧化锰氧化Cr(Ⅲ )的反应过程受Cr(Ⅲ )和Mn2 + 的扩散、氧化锰表面覆盖度等因素影响。拟合权函数的k和A值表明其反应先快后慢 ,且锰矿物类型不同 ,初始反应速率有差异 ;随着温度的升高 ,Cr(Ⅲ )的氧化速率增加。