Multi-models and belief functions for decision support systems

In our complex world, more and more decisions to be taken depend of many parameters. Usually the analysis of the problem to solve is based on a probabilistic framework, but sometimes, due to some uncertainty, this framework is not suitable. In this case belief functions can easily represent the lack of knowledge, and other types of uncertainty. But the main problem, namely the complexity of the involved system, remains. In this paper, a representation of the system based on a multimodels allows decision support systems to deal with complex and uncertain problems.