CISE Research Resources: Resources for Software Engineering Research

EIA-0224368 Edward B.AllenDavid A.Dampier；Thomas PhilipMissisippi State University：Resources for Software Engineering Research这个提案来自EPSCoR州，旨在增强软件工程研究的能力，旨在获得数据库服务器、配置管理系统和工作站，以支持以下项目：1.职业：高性能计算的开源软件评估，2.使用探路者网络对软件需求的感知进行建模，3.软件演进中的自动化步骤，以及4.科学仪器的综合信息筛选。第一个项目使用回溯性案例研究，通过构建和评估在历史项目或发布期间可能开发的模型来确定软件的质量。第二种方法提供了揭示涉众和开发人员对软件需求的不同看法的方法，从而根据小组之间的概念差异及早指示潜在的需求问题。第三，创建基于语义的切片和变更合并的自动化技术，以促进小增量的可靠交付，设计在增量变更演变时将增量变更合并到软件系统中的自动化方法。最后一个项目是生成图像，它开发了一套综合的算法和工具，用于分析科学仪器数据，以支持在线监测和过程控制。信息筛选方法包括用于光谱图像分类的人工神经网络、用于快速比较光谱的统计方法、基于规则的热成像模型和使用自主代理的模糊分类。

EIA-0224368Edward B. AllenDavid A. Dampier; Thomas PhilipMississippi State UniversityCISE RR: Resources for Software Engineering ResearchThis proposal from an EPSCoR state, enhancing capabilities for software-engineering research, aims at acquiring a data base server, a configuration management system, and workstations to support the following projects:1. CAREER: Assessment of Open-Source Software for High-Performance Computing,2. Using Pathfinder Networks to Model Perceptions of Software Requirements,3. Automating Steps in Software Evolution, and4. Integrated Information Sifting for Scientific Instruments.The first project ascertains the quality of the software using retrospective case studies by building and evaluating models that could have developed during the historical projects or releases. The second provides methods for uncovering divergent perceptions of software requirements by stakeholders and developers, thus giving an early indication of potential requirements issues in light of conceptual differences among groups. The third, creating automated techniques for semantics-based slicing and change-merging that will facilitate the reliable delivery of small increments, devises automated methods for incorporating incremental changes into software systems as they evolve. The last project, producing of imagery, develops an integrated set of algorithms and tools for analysis of scientific-instrument data to support on-line monitoring and process control. The information sifting approach includes artificial neural networks for classification of spectral images, statistical methods for rapid comparison of spectra, rule-based models of thermal imaging, and fuzzy classification using autonomous agents.