Integration of Internationally Distributed KB/DB on Material

本项目是一个通过计算机网络智能整合不同计算机资源进行材料设计的“虚拟实验室”的方案。这个虚拟实验室的范例是一个计算环境，我们可以通过计算机网络共享解决方案、材料设计、材料选择、工件维护、生命周期评估和许多未来问题的不同级别的信息。在这个项目中，材料设计的基本思想是为每个设计要求准备一个限定解决方案的窗口，最好能涵盖材料的复杂性和多样性。提出了几种方法，即第一性原理方法、结构图方法、功能图、机制图和由一套生产规则描述的材料开发经验序列。在整合的过程中，需要对材料信息的增长、合并和浏览进行快速有效的管理。此外，识别重要的外在关系的方法将成为提取每个设计问题的本质关系的必要条件。评估了所有可能为材料设计准备诱人的计算环境的可能性，并作为其核心节点，我们选择了一个用于预测不同材料电子结构的第一原理库和元素计算数据库，晶体数据库，势数据库，微观结构数据库和材料特性数据库。在原子间势的推导和评价活动中，说明了如何将不同结构层之间的数据和知识联系起来。该程序显示了不同知识库之间的相互作用，以阐明适当的抽象/参数化，以增加材料设计中的可预测性和计算效率。整合不同数据和知识之后的一个大问题是所收集信息的质量。为了解决这一问题，本文以时空浏览器为原型，讨论了如何通过重新缩放、重排和模式匹配获取知识进行分类，并作为系统部分实现。在处理动态微观结构变化时，这一问题变得非常严重，并对原型系统的时间和空间基本概念进行了分析和考虑。少

This project is a proposal to generate a "virtual laboratory" by showing an exemplar of intelligent integration of different computerized resources for materials design through computer network. This exemplar of virtual laboratory is an computational environment where we can share different levels of information for solution, materials design, materials selection, maintenance of artifacts, life cycle assessment and many future issues through computer networks.The basic idea of materials design in this project is to prepare windows which confine solutions for each design requirement and preferably covers the materials complexity and diversity. Several approaches have been proposed, namely, first principles approaches, structure map approaches, functional maps, mechanism maps and sequences of material development experiences described by a set of production rules. In the process of integration methods will be needed to quickly and efficiently manage the growth, merge and perusal of mater … More ials information. Furthermore, methods to identify extrinsic relations of significance will become mandatory to extract essential relations for each design problem.All the possibilities that may prepare seductive computational environments for materials design are evaluated and as a set of core nodes for it, we select a library for first principles and calculated data of elements to predict electronical structures for different materials, crystallographic data bases, potential databases, microstructural databases and materials properties databases. How to connect data and knowledge between different structural layrs are illustrated in the activity of deriving and evaluating of interatomic potentials. This procedure shows the interactions between different knowledge bases to articulate suitable abstraction/parameterization to increase a predictability as well as an efficiency of calculation in materials design.The big issue that comes after integration of different data and knowledge concerns the quality of collected information. In order to deal with this issue browsers with respect to time and space is developed as a prototype, where how to acquire knowledge through rescaling, rearrangement and pattern matching for classification is discussed and partly implemented as as system. This issue becomes serious in dealing with dynamic microstructural changes and basic concepts on time and space are analyzed and taken into accounts for a specification of the prototype system. Less