Hierarchical Strategy Learning with Hybrid Representations
Sung‐Wook Yoon, Subbarao Kambhampati · 2007
Good problem solving knowledge for real life domains is hard to define in a single representation. In some situations, a direct policy is a better choice while in others, value func-tion is better. Typically, direct policy representation is better suited to strategic level plans, while value function represen-tation is better suited to tactical level plans. We propose a hybrid hierarchical representation machine (HHRM) where direct policy representation and value function based repre-sentation can co-exist in a level-wise fashion. We provide simple learning and planning algorithms with our new rep-resentation and discuss their application to Airspace Decon-fliction domain. In our experiments, we provided our sys-tem LSP with two level HHRM for the domain. LSP could successfully learn from limited number of experts ’ solution traces and show superior performance compared to average of human novice learners.