Combining Two Fast-Learning Real-Time Search Algorithms Yields Even Faster Learning
David Furcy, Sven Koenig · 2014
Abstract. Real-time search methods, such as LRTA*, have been used to solve a wide variety of planning problems because they can make deci-sions fast and still converge to a minimum-cost plan if they solve the same planning task repeatedly. In this paper, we perform an empirical eval-uation of two existing variants of LRTA * that were developed to speed up its convergence, namely HLRTA * and FALCONS. Our experimental results demonstrate that these two real-time search methods have com-plementary strengths and can be combined. We call the new real-time search method eFALCONS and show that it converges with fewer actions to a minimum-cost plan than LRTA*, HLRTA*, and FALCONS. 1