Publications• Sorted by Date • Classified by Publication Type • Classified by Research Category • Influence-Augmented Online Planning for Complex EnvironmentsJinke He, Miguel Suau, and Frans A. Oliehoek. Influence-Augmented Online Planning for Complex Environments. In Advances in Neural Information Processing Systems 33, pp. 4392–4402, December 2020. DownloadAbstractHow can we plan efficiently in real time to control an agent in a complex environ- ment that may involve many other agents? While existing sample-based planners have enjoyed empirical success in large POMDPs, their performance heavily relies on a fast simulator. However, real-world scenarios are complex in nature and their simulators are often computationally demanding, which severely limits the performance of online planners. In this work, we propose influence-augmented online planning, a principled method to transform a factored simulator of the entire environment into a local simulator that samples only the state variables that are most relevant to the observation and reward of the planning agent and captures the incoming influence from the rest of the environment using machine learning methods. Our main experimental results show that planning on this less accurate but much faster local simulator with POMCP leads to higher real-time planning performance than planning on the simulator that models the entire environment. BibTeX Entry@inproceedings{He20NeurIPS, author = {He, Jinke and Suau, Miguel and Oliehoek, Frans A.}, title = {Influence-Augmented Online Planning for Complex Environments}, booktitle = NIPS33, year = 2020, month = dec, OPTeditor = {H. Larochelle and M. Ranzato and R. Hadsell and M. F. Balcan and H. Lin}, OPTpublisher = {Curran Associates, Inc.}, pages = {4392--4402}, url = {https://proceedings.neurips.cc/paper/2020/file/2e6d9c6052e99fcdfa61d9b9da273ca2-Paper.pdf}, keywords = {refereed}, abstract = { How can we plan efficiently in real time to control an agent in a complex environ- ment that may involve many other agents? While existing sample-based planners have enjoyed empirical success in large POMDPs, their performance heavily relies on a fast simulator. However, real-world scenarios are complex in nature and their simulators are often computationally demanding, which severely limits the performance of online planners. In this work, we propose influence-augmented online planning, a principled method to transform a factored simulator of the entire environment into a local simulator that samples only the state variables that are most relevant to the observation and reward of the planning agent and captures the incoming influence from the rest of the environment using machine learning methods. Our main experimental results show that planning on this less accurate but much faster local simulator with POMCP leads to higher real-time planning performance than planning on the simulator that models the entire environment. } }
Generated by
bib2html.pl
(written by Patrick Riley) on
Mon Oct 07, 2024 14:17:04 UTC |