Experience-based reasoning is a set of techniques encompassing instance-based, case-based, and analogy-based reasoning. All of these methods are useful when problem states can be readily represented, but effective control algorithms have not yet been found. In these cases, doing what worked well in the past may be an effective strategy. Experience-based methods do exactly that. They determine how similar the current state of the system is to a past state, and then apply the past state’s action. Although not immediately obvious, this reasoning methodology admits a great deal of creativity in the problem solving process. In particular, the determination of the most similar past state is an area of particular interest to me, as this is a key area in which contextual and goal-based knowledge can influence the control choices made.