基于知识的协同过程建模与实现技术研究

为实现协同过程中知识资源的优化利用,提出了基于知识的协同过程,给出了系统实现框架,研究了该系统的运行过程及相关技术,包括构建企业知识库、基于知识的协同过程的建模、运作过程中知识流与工作流的集成以及协同团队的动态评价。还简要介绍了企业间基于知识的协同过程。最后给出了一个具体应用的例子。     

工作流系统的异常处理

工作流执行时可能存在多种潜在的工作流异常,对这些异常进行预测并及时处理已经成为当前学者和软件开发商密切关注的课题。本文介绍了工作流系统的异常处理问题,综述了国内外的研究现状,明确了异常处理的概念和相关处理方法,并基于对现有研究方法的分析与评价,指出了工作流异常处理研究的发展方向。     

Graph theory based formulation of multi-period distribution expansion problems

In this paper a multi-period planning problem, with arbitrarily defined planning goals, is formulated in terms of graph theory. The proposed formulation represents a multi-period planning problem as a weighted graph problem and thus decomposes original problem into a number of sequences (spanning paths) of static planning problems without loss of accuracy. This graph problem is solved using dynamic programming technique. The proposed dynamic programming algorithm guaranties that optimal solution of multi-period planning problems will be found efficiently, assuming that optimality of static planning problems is guarantied. Detailed numerical results and comparisons presented in the paper show that proposed approach could improve noticeably the quality of multi-period solutions.     

Discrepancies between knowledge and practice of childhood motor vehicle occupant safety in Nova Scotia—A population-based study

Objectives

To determine discrepancies between knowledge and practice of childhood motor vehicle restraints (CMVRs) and vehicle seating position amongst parents within the province of Nova Scotia.

Design

Random telephone survey.

Setting

The Canadian province of Nova Scotia.

Subjects

Four hundred and twenty-six households with at least one child under the age of 12 years, totaling 723 children.

Main outcome measures

The proportion of parents whose children who should be in a specific stage of CMVR and sitting in the rear seat of the vehicle, and who demonstrate correct knowledge of that restraint system and seating position, yet do not use that restraint system/seating position for their child (demonstrate practice discrepant from their knowledge).

Results

Awareness of what restraint system to use is good (>80%). However, knowledge of when it is safe to graduate to the next stage is low (30–55%), most marked for when to use a seatbelt alone. Awareness of the importance of sitting in the rear seat of a vehicle was universal. Discrepancies between knowledge and practice were most marked with booster seats and rear-seating of older children. Factors influencing incorrect practice (prematurely graduated to a higher-level restraint system than what is appropriate for age and weight) included lower household income, caregiver education level, and knowledge of when to graduate from forward-facing car seats and booster seats. Incorrect practice was also more commonly observed amongst children of weight and/or age approaching (but not yet reaching) recommended graduation parameters of the appropriate CMVR.

Conclusions

Discrepancies between knowledge and practice are evident through all stages of CMVRs, but most marked with booster seats. The roles of lower socioeconomic status and gaps in CMVR legislation, in influencing discrepant practice, must be acknowledged and suggest the need for targeted education concurrent with development of comprehensive all-stages CMVR policies.     

基于结合面特性的机床整机精度设计系统研究

阐述了基于结合面特性的机床整机精度设计系统结构以及4大关键模块,并利用VB开发工具实现了基于结合面特性的机床整机精度设计系统。利用该系统实现了机床结合面特性参数的分析与管理,并基于多体系统理论建立了三轴数控机床静态几何精度模型,可对机床在整个加工空间的几何误差进行预估并对误差源进行敏感性分析;同时集成ANSYS软件,基于结合面特性可对机床整机动态性能进行分析。另外,通过有效管理机床设计过程中的设计经验以及动静态精度分析的知识,将有效地支持机床设计。     

Product portfolio map:a visual tool for supporting product variant discovery and structuring

A fundamental criterion for reusing and continuously improving knowledge in product development is ensuring that the knowledge is explicit and visual.This paper is based on the situation of an engineer-to-order(ETO) manufacturing company,where historically grown product variety and related knowledge are diffuse(tacit).Consequently,several resources are used in(re)developing derivatives of previous products rather than innovating new ones.To establish a more competitive configure-to-order(CTO) product strategy,product knowledge needs to be revealed,systemized,and structured,and thus made explicit.Hence,product-specific knowledge and product variants have been analyzed and subsequently mapped at architectural,functional,and physical levels in one unified map and tested in the form of a proof-of-concept(POC)demonstrator with the introduced SME company.The result is a product portfolio map that forms a base for defining a systemized,transparent,unified product variant overview,which can be used as a basis for implementing a cross-variant product architecture and supporting knowledge-based approaches.     

结合形状约束的Graph Cut行人分割

传统的Graph Cut算法没有对目标的形状予以限制,很难得到语义化的分割结果,即无法保证分割出来的是“行人”。针对该问题提出一种结合形状和底层特征的Graph Cut算法。对于行人分割,用大量真实行人轮廓来表达“行人”的先验形状,对Graph Cut分割算法予以约束,同时构建一个行人模板的层次树以减少匹配时间;并且提出一种区分性的外观模型来替换原来的外观模型。实验结果证明,该算法的分割结果明显优于传统Graph Cut算法的分割结果,所得到的轮廓与真实的行人轮廓比较吻合。     

无参数局部保持投影及人脸识别

局部保持投影(LPP)通过构造近邻图来保持样本的局部结构,在构图过程中,LPP面临复杂的参数选择问题.为解决此问题,提出无参数局部保持投影(PLPP)算法.首先设计一种无参数的构图方法,能够动态地获取样本的近邻点并配置相应的边权.其次,利用该构图方法,PLPP通过寻求最佳投影矩阵,用于保持样本在低维空间的局部结构.由于PLPP在构图过程中并未设置任何参数且采用余弦距离设置边权,因此PLPP计算更加方便快捷且对离群样本更具鲁棒性.另外,为进一步提升PLPP的识别性能,在PLPP的基础上通过加入样本的类别信息,提出监督的无参数局部保持投影算法(SPLPP).最后,在ORL、FERET及AR人脸库上的实验验证了PLPP与SPLPP的有效性.     

基于最小生成树的图像分割

基于最小生成树思想,给出了一种利用改进的最小生成树进行图像分割的方案,减少了最小生成树的构建过程,对初分割的结果利用NNG算法进行合并。该方案节约了分割时间,并且对分割后的图像进行了有效的合并,达到了较好的分割效果。     

贷款组合优化决策问题研究

针对贷款组合优化决策模型的求解问题,提出了用于求解该问题的知识进化算法,阐明了算法的具体实现过程,采用VB语言编写了相应的仿真实验程序。通过多次的仿真实验结果比较,表明了算法在计算时间、稳定性和寻优能力上的有效性。