Belief Change is the study of how an intelligent agent changes its beliefs according to the information it receives. There are three ways of changing one's beliefs, namely: adding a new belief to the body of beliefs, removing an existing belief from it, and adding a new belief while avoiding contradictions. The factors involved in belief change include beliefs of the agent, their relative strengths, their importance, relations between the beliefs. These four in particular jointly constitute the belief state of an agent. A realistic modeling of belief change should take into account two things: the real agents have limited resources and they perform belief changes more than once.
Belief change deals with the interaction of received information and the belief state over time. Most of the traditional approaches for belief change ignore the temporal aspects of belief change. They also ignore the real-world agents that are limited by memory, time to reach a decision or ability to reason. There is a need to develop a belief change theory that works for these resource-bounded agents.
Ideal rational change needs all possible inferences to be made from the given set of beliefs, using as much time as needed. Due to its limitations, an agent in the real world is incapable of performing ideal rational change. Instead, the agent approximates it. A real-world agent often performs successive changes to its belief state. It is necessary to study the effect of approximated belief change when the agent performs successive changes to its belief state. The purpose of my PhD thesis is exactly to do that.
