Filters








15 Hits in 4.1 sec

mGPT: A Probabilistic Planner Based on Heuristic Search

B. Bonet, H. Geffner
2005 The Journal of Artificial Intelligence Research  
This version, called mGPT, solves Markov Decision Processes specified in the PPDDL language by extracting and using different classes of lower bounds along with various heuristic-search algorithms.  ...  The heuristic-search algorithms use these lower bounds for focusing the updates and delivering a consistent value function over all states reachable from the initial state and the greedy policy.  ...  Acknowledgements mGPT was built upon a parser developed by John Asmuth from Rutgers University and Håkan Younes from Carnegie Mellon University. We also thank David E.  ... 
doi:10.1613/jair.1688 fatcat:amzieprqezgrlblyqdkindwmtq

Engineering a Conformant Probabilistic Planner

N. Onder, G. C. Whelan, L. Li
2006 The Journal of Artificial Intelligence Research  
We explain how we adapt distance based heuristics for use with probabilistic domains. Probapop also incorporates heuristics based on probability of success.  ...  We present a partial-order, conformant, probabilistic planner, Probapop which competed in the blind track of the Probabilistic Planning Competition in IPC-4.  ...  Acknowledgments This work has been supported by a Research Excellence Fund grant to Nilufer Onder from Michigan Technological University. We thank JAIR IPC-4 special track editor David E.  ... 
doi:10.1613/jair.1701 fatcat:wvrn55jg4zdl3ax7hlt5r2vtu4

Bridging the Gap Between Probabilistic Model Checking and Probabilistic Planning: Survey, Compilations, and Empirical Comparison

Michaela Klauck, Marcel Steinmetz, Jörg Hoffmann, Holger Hermanns
2020 The Journal of Artificial Intelligence Research  
We use this benchmark set as a basis for an extensive empirical comparison of various approaches from the model checking community, variants of value iteration, and MDP heuristic search algorithms developed  ...  On a per benchmark domain basis, techniques from one community can achieve state-ofthe-art performance in benchmarks of the other community.  ...  Related Work We presented a compilation from the probabilistic planning domain definition language (Younes & Littman, 2004) into the widely used model checking language Jani (Budde et al., 2017) .  ... 
doi:10.1613/jair.1.11595 fatcat:7kwnq5yx7fbn5lnvfvp72irvhy

Non-deterministic planning methods for automated web service composition

George Markou, Ioannis Refanidis
2015 Artificial intelligence research  
Particularly, since web services operate in a stochastic environment, their output is not predictable, and the problem is formulated as a non-deterministic planning one.  ...  This article presents a critical, comprehensive and up-to-date review of the literature concerning alternative non-deterministic planning methods, including probabilistic planning, determinization methods  ...  The search is based on a forward-chaining algorithm that exploits domain-specific search-control heuristics.  ... 
doi:10.5430/air.v5n1p14 fatcat:ph7cdzkc7vefni35ewcwc4inpq

Real-Time Path Planning using a Simulation-Based Markov Decision Process [chapter]

Munir Naveed, Andrew Crampton, Diane Kitchin, Lee McCluskey
2011 Research and Development in Intelligent Systems XXVIII  
Planners embedding each technique were applied to a typical RTS game and evaluated using the game score and the planning cost. The empirical evidence demonstrates the success of MCRT-planner.  ...  We have implemented the technique in MCRT-planner, a program which solves non-deterministic path planning problems in imperfect information RTS games, and evaluated it in comparison to four other state  ...  The current dynamic programming based planners such as mGPT [6] are applicable in domains that are modeled in a specific planning language (e.g.  ... 
doi:10.1007/978-1-4471-2318-7_3 dblp:conf/sgai/NaveedCKM11 fatcat:4tyhg6eqjjdijnjcbrqaf4opey

DeepSym: Deep Symbol Generation and Rule Learning from Unsupervised Continuous Robot Interaction for Planning [article]

Alper Ahmetoglu, M. Yunus Seker, Justus Piater, Erhan Oztop, Emre Ugur
2021 arXiv   pre-print
From its branches we extract probabilistic rules and represent them in PPDDL, allowing off-the-shelf planners to operate on the robot's sensorimotor experience.  ...  actions; and generated effective plans to achieve goals such as building towers from given cubes, balls, and cups using off-the-shelf probabilistic planners.  ...  For plan generation, the mGPT off-the-shelf probabilistic planner (Bonet and Geffner, 2005) with FF heuristic (Hoffmann and Nebel, 2001 ) was used.  ... 
arXiv:2012.02532v2 fatcat:wk6ia4eexvhofcilrb5zntoqh4

The First Probabilistic Track of the International Planning Competition

H.L.S. Younes, M. L. Littman, D. Weissman, J. Asmuth
2005 The Journal of Artificial Intelligence Research  
The 2004 International Planning Competition, IPC-4, included a probabilistic planning track for the first time.  ...  JAIR editor David Smith and his anonymous reviewers provided invaluable insights on the document that we tried to reflect in this final manuscript.  ...  This material is based upon work supported by the National Science Foundation under Grant No. 0315909 and the Royal Swedish Academy of Engineering Sciences (IVA) with grants from the Hans Werthén fund.  ... 
doi:10.1613/jair.1880 fatcat:q4uc4glzqngdtpwxfl2slmdcp4

Decision-Theoretic Planning with non-Markovian Rewards

S. Thiebaux, C. Gretton, J. Slaney, D. Price, F. Kabanza
2006 The Journal of Artificial Intelligence Research  
The current version of NMRDPP implements, under a single interface, a family of methods based on existing as well as new approaches which we describe in detail.  ...  A decision process in which rewards depend on history rather than merely on the current state is called a decision process with non-Markovian rewards (NMRDP).  ...  Simple Heuristic on the Number of Expanded States search informed by the above heuristic (LAO(h)).  ... 
doi:10.1613/jair.1676 fatcat:zf7e3w3yirhytcfwrx2hjvtdfe

Goal Probability Analysis in Probabilistic Planning: Exploring and Enhancing the State of the Art

Marcel Steinmetz, Jörg Hoffmann, Olivier Buffet
2016 The Journal of Artificial Intelligence Research  
We design and explore a large space of heuristic search algorithms, systematizing known algorithms and contributing several new algorithm variants.  ...  Our evaluation clarifies the state of the art, characterizes the behavior of a wide range of heuristic search algorithms, and demonstrates significant benefits of our new algorithm variants.  ...  We thank Christian Muise for his Probabilistic-PDDL extension of the FD parser. We thank Andrey Kolobov for discussions. We thank the anonymous reviewers, whose comments helped to improve the paper.  ... 
doi:10.1613/jair.5153 fatcat:klwoagf4c5dmfgq5ifzaxebbou

Practical solution techniques for first-order MDPs

Scott Sanner, Craig Boutilier
2009 Artificial Intelligence  
To demonstrate the applicability of our techniques, we present proof-of-concept results of our first-order approximate linear programming (FOALP) planner on problems from the probabilistic track of the  ...  ., those stated in the probabilistic planning domain description language, or PPDDL) ground the specification with respect to a specific instantiation of domain objects and apply a solution approach directly  ...  We compared FOALP to the three other top-performing planners on these problems: NMRDPP is a temporal logic planner with human-coded control knowledge [77] ; mGPT is an RTDP-based planner [10] ; (Purdue  ... 
doi:10.1016/j.artint.2008.11.003 fatcat:6l2ocjejfrgixdgwqscblhadni

From Skills to Symbols: Learning Symbolic Representations for Abstract High-Level Planning

George Konidaris, Leslie Pack Kaelbling, Tomas Lozano-Perez
2018 The Journal of Artificial Intelligence Research  
off-the-shelf planner.  ...  In addition, we show that learning the relevant probability distributions corresponds to specific instances of probabilistic density estimation and probabilistic classification.  ...  George Konidaris was partially supported during this work by an Intelligence Initiative Fellowship from MIT, and by a DARPA Young Faculty Award under agreement number D15AP00104.  ... 
doi:10.1613/jair.5575 fatcat:sjchgaf6m5a4tn7icvz7p6p3q4

Model-driven engineering of planning and optimisation algorithms for pervasive computing environments

Anthony Harrington, Vinny Cahill
2011 2011 IEEE International Conference on Pervasive Computing and Communications (PerCom)  
We present an empirical evaluation of the impact of our approach on the development effort associated with a pervasive computing application from the Intelligent Transportation Systems (ITS) domain, and  ...  We present a layered domain model containing a set of objectoriented specifications for modelling physical and sensor/actuator infrastructure and state-space information.  ...  Automated planners take as input a description of the problem to be solved and produce as output a plan to govern the actions taken by an application.  ... 
doi:10.1109/percom.2011.5767582 dblp:conf/percom/HarringtonC11 fatcat:xws7lfxfurhpzkurjr3mvblsqy

Model-driven engineering of planning and optimisation algorithms for pervasive computing environments

Anthony Harrington, Vinny Cahill
2011 Pervasive and Mobile Computing  
We present an empirical evaluation of the impact of our approach on the development effort associated with a pervasive computing application from the Intelligent Transportation Systems (ITS) domain, and  ...  We present a layered domain model containing a set of objectoriented specifications for modelling physical and sensor/actuator infrastructure and state-space information.  ...  Automated planners take as input a description of the problem to be solved and produce as output a plan to govern the actions taken by an application.  ... 
doi:10.1016/j.pmcj.2011.09.005 fatcat:33kdqls6vrgzxdzv53xjalntk4

Search-Based Task Planning with Learned Skill Effect Models for Lifelong Robotic Manipulation [article]

Jacky Liang, Mohit Sharma, Alex LaGrassa, Shivam Vats, Saumya Saxena, Oliver Kroemer
2021
Experiments demonstrate the ability of our planner to integrate new skills in a lifelong manner, finding new task strategies with lower costs in both train and test tasks.  ...  Prior works on planning with skills often make assumptions on the structure of skills and tasks, like subgoal skills, shared skill implementations, or learning task-specific plan skeletons, that limit  ...  The authors thank Kevin Zhang for assistance on real-world experiments. This work is supported by NSF Grants No. DGE 1745016, IIS-1956163, and CMMI-1925130, the ONR Grant No.  ... 
doi:10.48550/arxiv.2109.08771 fatcat:aocp3xkebjhu3dzgdtcrrdgt5q

JOBPLAN-A NEW INTEGRATED REPRESENTATION AND PLANNER FOR BATCH JOB WORKFLOW AUTOMATION JOBPLAN-A NEW INTEGRATED REPRESENTATION AND PLANNER FOR BATCH JOB WORKFLOW AUTOMATION

Tracey Lall, Tracey Lall
2011 unpublished
The implementation and evaluation of a prototype planner "JobPlan" on key domain scenarios illustrating these features is presented. ii Acknowledgements Matthew Stone for his invaluable guidance, insight  ...  Plans and beliefs are represented as a workflow state machine governed by a clearly defined dynamics. Time based goals are handled by treating time as a fluent.  ...  of a probabilistic event.  ... 
fatcat:klaxrui5urhj7ca5ejhc7nvmuy