Abstract

It is well known that the computational complexity of propositional knowledge base revision is at the second level of polynomial hierarchy. A way to solve this kind of problems is to introduce approximate algorithms. In this paper, an approximate approach is introduced for belief change. Operators, which satisfy the AGM rational postulates, are defined to change belief sets or belief bases. Furthermore, approximate algorithms to implement the revision of finite belief bases are presented. The time complexities of the approximate algorithms shown in this paper are at lower level than the time complexities of the existed approaches in literatures, although they may not generate the optimal solution, and this is meaningful from the theoretical point of view.