Reasoning with Imprecise Context Using Improved Dempster-Shafer Theory

In pervasive computing environment the contexts are usually imprecise and incomplete due to unreliable connectivity, user mobility, and resource constraints. In this paper we present an approach based on the Dempster-Shafer Theory (DST) for the reasoning with imprecise context. To solve the two fundamental issues of the DST, computation intensiveness and the Zadeh paradox, we filer out excrescent subsets based on their energy to reduce the number of subsets, and employ the concept of evidence loss and approval degree of evidence in the combining process.