Memory and planning processes in solutions to well-structured problems

Although many studies in the problem-solving literature have considered the factors that might determine the strategies that are employed to solve well-structured problems, these have typically focused upon variants of means-end analysis. In general, such models imply that strategies unfold in a temporally forward direction, that problem solvers typically restrict forward-planning activities to just one or two moves ahead of the current problem state, and that one important heuristic is the avoidance of previous moves. Although studies have demonstrated the importance of such anti-looping heuristics, few have systematically explored the possibility that problem solvers may also plan retrospectively in order to try and assess whether a move might take them back to a state that they have previously visited. Those models of problem solving that promote the role of an anti-looping heuristic have assumed that the ability to use such a heuristic is based upon memory for previous states, but other interpretations are possible. In this paper several studies are reported that attempt systematically to explore participants' attempts to recognize previously visited problem-solving states. The findings suggest that there is a systematic relationship between the success of this process, the time taken to make this judgement, and distance from the current state. It is also demonstrated that estimations about where future positions are likely to occur are symmetrical to estimations about past positions. It is suggested that this provides evidence that problem solvers engage in retrospective planning processes in order to try and avoid previous moves, and that this strategy may not be based straightforwardly upon their ability to remember previous problem states.