Realtime Search Performance
This chapter investigates available real-time search algorithms, e.g., Real-Time-A* (RTA*), Learning-Real-Time-A* (LRTA*) [Korf, 1990] and Local-Consistency-Maintenance (LCM) [Pemberton and Korf, 1992]. Though realtime search algorithms have some learning capability, previous research has been focused on the performance of each problem solving trial. For example, LRTA* learns the exact cost to the goal along the optimal path to the goal. However, there is almost no research on the learning efficiency of realtime search. This chapter is intended to evaluate the learning process to clarify the following three basic questions [Mizuno and Ishida, 1995].
KeywordsActual Cost Goal State Optimal Decision Problem Solver Problem Space
Unable to display preview. Download preview PDF.