Model Composition in Multi-dimensional Data Spaces

Model composition is an important problem in model management. In this paper, we propose a new method to support model composition in multi-dimensional data spaces. We define a model as a 6- tuple with input interface and output interface. An algorithm for model composition and execution is given. Moreover, the method has been applied into a practical project. The running statistics showed that there had been 105 instances of model composition, and 89 decision problems had been effectively solved.