AI Planning under uncertainty

            
             "It is the mark of an instructed mind to rest satisfied with that degree of precision which the nature of the subject admits, and not seek exactness where only an approximation of the truth is possible."".
             Aristotle.
             Introduction.
             The aim of this essay is to describe the role that uncertainty plays within the field of AI planning. This essay is aimed at undergraduate computer scientists with little previous knowledge of the topic. To describe this field it is necessary to give a brief description of what AI planning is and what it is used for, and then to describe what is meant by uncertainty and how this affects our plans. I will then go on to talk about ways of incorporating uncertainty into plans, and planning systems, and how this can be represented. Finally I will talk about the future role of uncertainty in AI planning.
             What is planning?.
             Planning is the process of finding a sequence of steps which when put into action will achieve an intended result. In our everyday lives this is usually a simple process; we plan our actions all the time, whether it's deciding to go to the bank before going to the pub so that we can afford to buy a round of drinks, or following a recipe to make a Christmas pudding. Planning seems such a natural process to us that we take it for granted, yet when the process by which we plan is analysed, it turns out to be trickier to reproduce than it seems at first glance.
             AI planning.
             AI planning is the body of knowledge which is concerned with reproducing the planning process; this usually means the creation of planning systems. A planning system is fairly intuitively, a system which produces plans, or ordered sequences of individual actions. The creation of an ordered plan requires creating some kind of world model which includes appropriate objects and actions, that can be used to complete a desired goal or state. Each action will usually have a set of restrictions which need to be met before performing that act, and a set of outcomes which come about through having performed that act, or pre-conditions and effects.