基于关联假设的定性组合建模算法

组合建模是基于建模假设和模型片断库，构造定性仿真模型的自动化方法。关联假设是一类模型假设，它既可以用来定义模型片断本身，也可以定义有关模型片断使用的决策信息，将模型库与建模任务联系起来。在知识分类的基础上，该文提出了基于关联假设的模型组合算法，解决了定性仿真模型的构造问题，主要内容包括关联性的表示、关联推理算法及其具体实现，最后给出一个实例。

Algorithm of Qualitative Compositional Modeling Based on Relevance Assumption

Based on modeling assumption and model fragment library, compositional modeling is the method for building models of qualitative simulation automatically. Relevance assumption is one kind of modeling assumption that can be used to define model fragments themselves as well as decision information's definition about management of model fragments. It relates model fragment library to the current task of qualitative modeling. After making a clear classification of model knowledge, this paper proposes a relevance-assumption-based model composing algorithm, partially solves the puzzle of qualitative model construction. The main research items include relevance representation, reasoning algorithms and its implementation. In the end, a simple example is given to testify the concrete procedure of this algorithm.