An open-source data manager for network models
An open-source data manager for network models
复制标题
DOI：
10.1016/j.envsoft.2019.104538
复制
发表时间：
2019-12
期刊：
影响因子：
--
通讯作者：
S. Knox;J. Tomlinson;J. Harou;P. Meier;D. Rosenberg;J. Lund;D. E. Rheinheimer
中科院分区：
文献类型：
--
作者：
S. Knox;J. Tomlinson;J. Harou;P. Meier;D. Rosenberg;J. Lund;D. E. Rheinheimer
为水或能源系统等资源网络开发模拟和优化模型越来越多地涉及到集成多个数据源和软件。连接多个模型并管理不同分析师组访问的数据是一项软件挑战。许多资源系统在计算机模型中被表示为节点和链接的网络，由一系列目标和规则驱动。我们提出了一个用Python编写的数据存储平台，它利用网络表示的通用性在单个部署中存储多个模型类型的数据。这个开源平台使用一致的数据格式为多个模型提供了一个通用的数据源，与基于文件的数据管理相比，减少了错误的可能性。当部署为Web服务时，它允许授权用户通过互联网安全地共享数据，促进协作。一个案例研究描述了一个水设施规划模型的托管，并附有一个工作实例。
Developing simulation and optimisation models for resource networks like water or energy systems increasingly involves integrating multiple data sources and software. Connecting multiple models and managing data accessed by different groups of analysts is a software challenge. Many resource systems are represented in computer models as networks of nodes and links, driven by a range of objectives and rules. We present a data storage platform, written in Python, which exploits the commonality of network representations to store data for multiple model types within a single deployment. This open-source platform provides a common source of data to multiple models using consistent data formats, reducing likelihood of error compared to file based data management. When deployed as a web service, it allows data to be shared securely among authorised users over the internet, facilitating collaboration. A case study describes the hosting of a water utility planning model, with an accompanying worked example.