Relaxation Optimizing Processes in Extended Probabilistic Space
T. Horiucliil · 1995
Classacal probabalastac relaxation melhod has been wadely used for solvang optamezntaon problems in various fields, including image processzng and pattern recognetaon. However we realaze that there exasts cases an which a probabalaty theoretac model as not adequate, especaally there exzsts zncompleteness an avaalable znformation by noses. In order to solve the problem, thas paper proposes a relaxation matchang method based on Dempster-Shafer theory. Then the update process an probabzlastac relaxaiaon method is deraved as a specaal case of Dempster's combznatzon rule in DS theory.