Approach for Performance Predicting of IDFA
Bai Zhi-jiang, Lixin Song · Jisuanji yingyong yanjiu · 2006
Analyze the IDFA~*(Iterative-based Depth First A~*) algorithm with an admissible and consistent heuristic function to predict the running time of the DFIA~* accurately.Traditional analysis use accuracy as character of the heuristic function and the function's effect is to decrease the actual branching factor.Our analysis employs the probability of heuristic values as character of the heuristic function and the function's effect is to reduce the actual search depth.The experimental result shows that the performance of DFIA~* on actual problems like Rubik's Cube can be accurately predicted with our main analytic conclusion.