ITR-(ASE+NHS)-(dmc+sim): Simulation Transformation for Dynamic, Data-Driven Application Systems (DDDAS)

Brogan，David0426971我们的世界越来越多地配备了仪器，并与遥感和计算设备联网，为增强分析和预测创造了新的机会。利用这些机会需要动态的、数据驱动的应用系统(DDDAS)，该系统由传感器、网络、计算资源、模拟软件和人在环中的复杂组合组成。该项目将开发一种端到端的DDDAS模拟设计方法和转换过程，以提高模拟适应运行时计算和数据资源的动态组合的潜力。该项目将部署一种称为.COERCE的横切方法来进行模拟设计和转换。在现有的动态、数据驱动的应用系统中，让模拟最有效地工作是迫在眉睫的挑战。强迫既是强制性(模拟的一种属性)，也是强制性(一种过程)。Coerciability解决了这样的实践和方法，设计者通过这些实践和方法获取关于如何转换模拟的知识(可能是在模拟语言构造中)，以支持未来的强制。凝聚性提供了模拟属性的捕获和表示。强制是使用人工控制的代码修改和半自动分析和优化的混合来转换模拟的过程，以满足一组特定的要求，可能与最初设计模拟的要求不同。通过强制，先前存在的(遗留)模拟可以被转换以适应新的或缺失的数据源，改变保真度，并改变它们在更大的数据驱动系统中的角色。

Brogan, David0426971Our world is increasingly instrumented and networked with remote sensing and computational devices, creating new opportunities for enhanced analysis and prediction. Exploitation of these opportunities requires dynamic, data-driven application systems (DDDAS) composed from complex combinations of sensors, networks, computational resources, simulation software, and humans-in-the-loop. The project will develop an end-to-end DDDAS simulation design methodology and transformation process that will improve the potential for simulations to adapt to the dynamic composition of computational and data resources at run time. The project will deploy a crosscutting approach to simulation design and transformation called .COERCE.. COERCE addresses the immediate challenges of making simulations work most effectively in existing dynamic, data-driven application systems. COERCE is both coercibility (a property of simulations) and coercion (a process). Coercibility addresses the practices and methods whereby designer knowledge about how a simulation can be transformed is captured (possibly in simulation language constructs) to support future coercion. Coercibility provides for the capture and representation of simulation properties. Coercion is the process of transforming a simulation using a human-controlled mix of code modification and semi-automated analysis and optimization to meet a specific set of requirements, possibly differing from those for which the simulation was originally designed. Through coercion, preexisting (legacy) simulations can be transformed to adapt to new or missing data sources, to change fidelity, and to change their roles in a larger data-driven system.