Garp3-Workbench for qualitative modelling and simulation
Garp3-Workbench for qualitative modelling and simulation
复制标题
DOI：
10.1016/j.ecoinf.2009.09.009
复制
发表时间：
2009-11-01
影响因子：
5.1
通讯作者：
Liem, Jochem
中科院分区：
文献类型：
--
作者：
Bredeweg, Bert;Linnebank, Floris;Liem, Jochem
易于使用的定性推理和建模工作台几乎不存在。这限制了这种人工智能技术的使用及其被更大受众的吸收。我们提出了Garp3，一个用户友好的工作台，允许建模者建立，模拟和检查系统行为的定性模型。该工作台采用图形表示，供用户与模型内容和仿真结果进行交互，并提供不同使用模式之间的无缝互操作性。领域专家可以使用Garp3在数字信息稀疏或不可用的情况下创建概念模型，或者当他们想要形式化他们对系统行为的概念理解时。Garp 3可应用于利益相关者管理或传播活动，以说明和解释现象，并促进参与者之间的讨论。工作台也可以用于正规教育，让学习者表达概念，或与现有模型交互，并支持他们发展对“事物如何工作”的理解。本文的主要目标是介绍这些技术的表示和推理方法，因为它们已经在Garp3工作台中开发和微调。重点在于有效生成系统行为状态图所需的符号非数值计算，同时考虑到用户需要能够跟踪和理解这种推理，无论是最终结果还是它提供的中间结果。(C)2009 Elsevier B.V.保留所有权利。
Easy to use workbenches for Qualitative Reasoning and modelling have been virtually nonexistent. This has a limiting effect on the use of this Artificial Intelligence technology and its uptake by a larger audience. We present Garp3, a user-friendly workbench that allows modellers to build, simulate, and inspect qualitative models of system behaviour. The workbench employs diagrammatic representations for users to interact with model content and simulation results, and provides seamless interoperability between the different modes of use. Domain experts can use Garp3 to create conceptual models in situations where numerical information is sparse or unavailable, or when they want to formalise their conceptual understanding of how systems behave. Garp3 can be applied to stakeholder management or dissemination activities to illustrate and explain phenomena, and facilitate discussion among participants. The workbench can also be used in formal education to have learners express concepts, or interact with existing models, and support them in developing their understanding of 'how things work'.Garp3 incorporates a range of techniques from Artificial Intelligence known as knowledge-based techniques. The main goal of this paper is to present the representation and reasoning methods of these techniques as they have been developed and fine-tuned within the Garp3 workbench. The focus hereby is on the symbolic, non-numerical calculations that are required to generate the state-graph of a system's behaviour efficiently, while taking into account that users need to be able to track and understand this reasoning, both in terms of the end result and the intermediate results it delivers. (C) 2009 Elsevier B.V. All rights reserved.