Zhengzhu Feng

Learn More
We describe an approach for exploiting structure in Markov Decision Processes with continuous state variables. At each step of the dynamic programming, the state space is dynamically partitioned into regions where the value function is the same throughout the region. We first describe the algorithm for piecewise constant representations. We then extend it(More)
We present a major improvement to the incremental pruning algorithm for solving partially observable Markov decision processes. Our technique targets the cross-sum step of the dynamic programming (DP) update, a key source of complexity in POMDP algorithms. Instead of reasoning about the whole belief space when pruning the cross-sums, our algorithm divides(More)
We present a simple, yet effective improvement to the dynamic programming algorithm for solving partially observable Markov decision processes. The technique targets the vector pruning operation during the maximization step, a key source of complexity in POMDP algorithms. We identify two types of structures in the belief space and exploit them to reduce(More)