遗传算法实现农村树状供水管网的优化设计

为改善农村树状供水管网的管径设计方案,应用数学建模和最优化理论,提出在确定管网布置方案的基础上,采用整数编码的遗传算法,建立管径组合方案优化模型。以经济性作为目标函数,并结合实例,编制相应计算程序,对某一农村供水管网进行优化设计,与传统设计方案进行对比。结果表明,该优化设计方案在经济性和水力特性都优于传统的设计方案。     

基于二元决策图的故障树分析方法

传统故障树分析的二元决策图方法受变量排序影响很大,可能使运算量和运行空间过大。为此,提出在故障树预处理的基础上,使用直接生成法以简化二元决策图的生成过程。该方法避免了变量排序的影响,提高了复杂故障树的分析效率。计算中首先通过结构化简和线性时间算法模块化完成预处理,然后将直接生成法应用于各个模块进而应用到整个故障树。最后将此方法与模块连接法、相邻底事件优先法进行实例比较,证明其生成二元决策图的规模更小且运算量更少。     

Pricing structures for long-term international contracts

Devising a satisfactory commodity pricing mechanism for incorporation into a 20–25 year supplier-customer contract poses many challenges. The authors compare some advantages and drawbacks of possible mechanisms, drawing partly on a relevance tree approach.     

事故树分析方法在水利水电工程建设生态风险评价中的应用研究

水利水电工程建设将会对河流生态系统产生一定影响,从而改变其系统内部的稳定性。为定性和定量研究水利水电工程建设生态风险,本文引入了事故树分析方法,利用该研究方法,构建了水利水电工程建设生态风险事故树,揭示了生态系统内部各关联因子之间的关系,从而对生态系统发生破坏的风险进行评价。     

泵站加压式树状供水管网优化设计研究

针对传统供水管网优化设计时未考虑整个管网的首末水头约束问题,提出了以管网系统年折算费用最小为目标函数,以水泵扬程和具有标准管径的干、支管管道长度为决策变量的泵站加压式树状供水管网优化设计数学模型。在该模型中,假定水泵扬程已知,将管网优化问题转化成管网投资线性子模型问题,并采用大系统分解-动态聚合方法对该子模型进行求解。研究结果表明,该方法寻优能力强,且计算时间短,计算结果精度较高,为泵站加压式树状供水管网系统优化设计提供了一种可行而有效的新方法。     

中美海绵城市建设中的树木管理比较

针对我国在海绵城市建设中未能足够重视城市树木对雨水径流削减功能的问题,在简述中美两国关于城市树木和径流削减关系研究的基础上,比较分析了两国城市树木管理方面的法规、专项规划以及规范标准中涉及雨水管理的相关内容,指出我国目前还存在缺乏对城市下垫面及城市树冠的精细化分析、对城市树冠的生态功能及相关指标重视不足、在树种和种植位置的选择上缺乏对雨水径流控制的考量等问题。建议加强基础资料的积累和公开、深化树木削减雨水径流能力的基础研究以及完善相关规划和规范标准。     

基于分类与预测的两段式就业指导系统的分析

本文研究分类与预测技术在学生就业指导中的应用。主要分为入学后的就业指导与求职前的就业指导。利用了决策树模型分析出贡献度最大的分裂属性,对学生进行入学前就业指导。利用基于类平均距离对学生可能就业的单位进行预测,为学生求职时提供有价值的指导信息。     

Minimum cost spanning tree problems with indifferent agents

We consider an extension of minimum cost spanning tree (mcst) problems in which some agents do not need to be connected to the source, but might reduce the cost of others to do so. Even if the cost usually cannot be computed in polynomial time, we extend the characterization of the Kar solution (Kar, 2002) for classic mcst problems. It is obtained by adapting the Equal treatment property: if the cost of the edge between two agents changes, their cost shares are affected in the same manner if they have the same demand. If not, their changes are proportional to each other. We obtain a family of weighted Shapley values. Three interesting solutions in that family are characterized using stability, fairness and manipulation-proofness properties.     

基于决策树算法SLIQ的应用实例

SLIQ算法是一种高速可调节的数据挖掘分类算法。着重解决当训练集数据量巨大,无法全部放入内存时,如何高速准确地生成判定树。文中把该算法用到邮政速递的客户关系管理系统中,对速递大客户进行分类分析,识别优质客户和潜在客户,获取的知识对邮政速递的决策有重要现实意义。     

Model to evaluate the cost-effectiveness of different antibiotics in the management of community-acquired pneumonia in Germany

Summary

This study investigated the primary cost-drivers and determinants of the cost-effectiveness of antibacterial treatment of community-acquired pneumonia (CAP) in Germany. It assessed the health care costs and consequences related to treatment initiated in the community using macrolides, fluoroquinolones, and cephalosporins. Patients were categorised according to disease severity. Decision analysis was used to consider the clinical and economic consequences of various treatment options from first-line treatment initiated by a primary care physician in the community to success or failure after third-line treatment in hospital.

The key cost drivers were the clinical success/failure rates of first-line treatment and, in moderate CAP, the cost per day of hospitalisation. Thus, antibiotics with the cheapest purchase price are not necessarily the most cost-effective treatment. This is because the extra costs associated with treatment failure, especially in more severely ill patients, can be much greater than the acquisition costs of the first-line antibiotic. Of the antibiotics considered, none was consistently found to be the most cost-effective across the full range of scenarios investigated. However, in moderate CAP the fluoroquinolones and cephalosporins were generally more cost-effective than the macrolides. Given the high cost associated with hospitalisation, and the low mortality rates in low-risk patients with mild and moderate CAP, successful outcomes and health care cost savings can be achieved if such patients receive appropriate antibiotics as first-line treatment in the community.