Serious Flaws in Korf et al.'s Analysis on Time Complexity of A*

Understanding the effect of a heuristic on the time complexity of the well-known A* search algorithm has been a subject of a large research body in AI. Among those researches is Korf et al.'s analysis (Korf & Reid, 1998; Korf et al., 2001; Korf, 2000a; Korf, 2000b), which concluded that the effect of a heuristic is to reduce the effective depth rather than effective branching factor. Since this conclusion conflicts with both previous analysis and recent analysis by other authors, we try to investigate the causes of such contradictions and finally find out that Korf et al.'s analysis has many serious flaws from which their main conclusion was drawn.