Study on self-adaptive fusion model with DST and PCR5

Classical conflict is used to express conflict degree in DST. But many studies show that it is not efficient in measuring conflict, because it only consider non-inclusive between evidences. To solve this problem, a new modified measure method of evidence conflict is proposed, which is based on conflict coefficient K and relative coefficient between evidences. It not only reflects the non-inclusive, but also reflects the differences between evidences. Then this paper gives two kinds of self-adaptive fusion model on the basis of DST and PCR5. It is proved by simulation that they are effective in dealing with conflict information. They converge quickly and help the system make the right decision logically. At last, comparison and analysis have be done between two models and some advices are given.