The Dempster-Shafer theory of evidence reasoning (D-S theory) has been
widely discussed and used recently, because it is a reasonable, conve
nient, and promising method to combine uncertain information from disp
arate sources with different levels of abstraction. On the other hand,
the D-S theory has sparked considerable debate among statisticians an
d knowledge engineers. The theory has been criticized and debated upon
its behavior and attributes, such as high computational complexity, e
vidence independency requirement in its combination rule, etc. some pr
incipal problems of the D-S theory are discussed in the paper. The rel
ationship of the D-S theory and the classical probability theory is an
alyzed first, and then a generalized evidence combination formula rela
xing the requirement of evidence independency is presented, which make
s the D-S theory more realistic to applications.