Z
Zhengzhu Feng
Researcher at University of Massachusetts Amherst
Publications - 12
Citations - 520
Zhengzhu Feng is an academic researcher from University of Massachusetts Amherst. The author has contributed to research in topics: Markov decision process & Partially observable Markov decision process. The author has an hindex of 9, co-authored 12 publications receiving 515 citations.
Papers
More filters
Proceedings ArticleDOI
Symbolic heuristic search for factored Markov decision processes
Zhengzhu Feng,Eric A. Hansen +1 more
TL;DR: A plnning algorithm is described that integrates two approaches to solving Markov decision processes with large state spaces in a novel way that exploits symbolic model-checking techniques and demonstrates their usefulness for decision-theoretic planning.
Proceedings ArticleDOI
Dynamic programming for structured continuous Markov decision problems
TL;DR: This work describes an approach for exploiting structure in Markov Decision Processes with continuous state variables and extends it to piecewise constant representations, using techniques from POMDPs to represent and reason about linear surfaces efficiently.
Posted Content
Dynamic Programming for Structured Continuous Markov Decision Problems
TL;DR: In this article, the state space is dynamically partitioned into regions where the value function is the same throughout the region, where the state variables can be expressed by piecewise constant representations.
Book ChapterDOI
Adaptive peer selection
TL;DR: This work denotes as peer selection the entire process of switching among peers and finally settling on one, and uses the methodology of machine learning for the construction of good peer selection strategies from past experience.
Proceedings ArticleDOI
Region-based incremental pruning for POMDPs
Zhengzhu Feng,Shlomo Zilberstein +1 more
TL;DR: In this paper, an incremental pruning algorithm for solving partially observable Markov decision processes (POMDPs) is proposed, where instead of reasoning about the whole belief space when pruning the cross-sums, the algorithm divides the belief space into smaller regions and performs independent pruning in each region.