On the Approximate Solution of POMDP and the Near-Optimality of Finite-State Controllers
|
|
- Eustacia Woods
- 5 years ago
- Views:
Transcription
1 On the Approximate Solution of POMDP and the Near-Optimality of Finite-State Controllers Huizhen (Janey) Yu Dimitri Bertsekas Lab for Information and Decision Systems, M.I.T. April
2 POMDP (Partially Observable Markov Decision Problems) 2 Very important in control and other fields
3 POMDP (Partially Observable Markov Decision Problems) 2 Very important in control and other fields Hard (because of infinite state space - the space of beliefs or conditional state distributions)
4 POMDP (Partially Observable Markov Decision Problems) 2 Very important in control and other fields Hard (because of infinite state space - the space of beliefs or conditional state distributions) Last frontier for Exact DP
5 POMDP (Partially Observable Markov Decision Problems) 2 Very important in control and other fields Hard (because of infinite state space - the space of beliefs or conditional state distributions) Last frontier for Exact DP Great challenge for Approximate DP
6 Approximation/Finite Parametrization Methods for POMDP 3 Let s focus on finite parametrizations of finite-state POMDP: Discretization/interpolation in belief space
7 Approximation/Finite Parametrization Methods for POMDP 3 Let s focus on finite parametrizations of finite-state POMDP: Discretization/interpolation in belief space Problem approximation (use a finite-state MDP to approximate a POMDP/aggregation)
8 Approximation/Finite Parametrization Methods for POMDP 3 Let s focus on finite parametrizations of finite-state POMDP: Discretization/interpolation in belief space Problem approximation (use a finite-state MDP to approximate a POMDP/aggregation) Piecewise linear approximations (approximate value and policy iteration)
9 Approximation/Finite Parametrization Methods for POMDP 3 Let s focus on finite parametrizations of finite-state POMDP: Discretization/interpolation in belief space Problem approximation (use a finite-state MDP to approximate a POMDP/aggregation) Piecewise linear approximations (approximate value and policy iteration) Optimization over a set of finite-state controllers
10 4 Outline of the Talk Introduction to POMDP (finite-state average cost) Lower bounds/mdp approximations to POMDP optimal cost function (a brief summary of our work) On Near-Optimality of the Set of Finite-State Controllers for Average Cost POMDP
11 5 References H. Yu, Approximate Solution Methods for Partially Observable Markov and Semi-Markov Decision Processes, Ph.D. Thesis, Dept of EECS, MIT, Jan H. Yu and D. P. Bertsekas, Discretized Approximations for POMDP with Average Cost, The 20th Conference on Uncertainty in Artificial Intelligence, 2004, Banff, Canada. H. Yu and D. P. Bertsekas, Near-Optimality of Finite-State Controllers, LIDS Report 2689, MIT, April 2006.
12 6 Introduction to POMDP Graphical model of discrete-time POMDP: S 0 u 0 S 1 S 2 Y 1 u 1 Y 2 S state space, Y observation space, U control space g(s, u) ξ π Π per-stage cost initial distribution of state history-dependent randomized policies π = {µ 0, µ 1,..., } µ 0 ( ), µ t (h t, ) P(U) h t = (u 0, y 1, u 1,..., y t ), observed history up to time t ξ, π stochastic process {S 0, U 0, S 1, Y 1, U 1,...} joint probability distribution P ξ,π
13 7 Expected Cost Criteria k-stage cost J π k Average cost (ξ) = EPξ,π k 1 g(s t, U t ) t=0 Jk (ξ) = inf π Π Jπ k (ξ) J π (ξ) = lim inf k J (ξ) = inf π Π Jπ (ξ) 1 k Jπ k (ξ) 1 Jπ + (ξ) = lim sup k k Jπ k (ξ) (ξ) = inf J + π Π Jπ + (ξ)
14 8 Important Analytical Features Sufficient statistic for control is ξ t : belief state/conditional distribution of the state
15 8 Important Analytical Features Sufficient statistic for control is ξ t : belief state/conditional distribution of the state k-stage policy cost function: linear in ξ
16 8 Important Analytical Features Sufficient statistic for control is ξ t : belief state/conditional distribution of the state k-stage policy cost function: linear in ξ Lower k-stage optimal cost function: piecewise linear/concave in ξ
17 8 Important Analytical Features Sufficient statistic for control is ξ t : belief state/conditional distribution of the state k-stage policy cost function: linear in ξ Lower k-stage optimal cost function: piecewise linear/concave in ξ Key questions for average cost: are the lower or upper optimal cost functions flat (constant/independent of the initial belief ξ)? Are they equal?
18 8 Important Analytical Features Sufficient statistic for control is ξ t : belief state/conditional distribution of the state k-stage policy cost function: linear in ξ Lower k-stage optimal cost function: piecewise linear/concave in ξ Key questions for average cost: are the lower or upper optimal cost functions flat (constant/independent of the initial belief ξ)? Are they equal? In POMDP constant cost across beliefs is far less likely than in MDP
19 9 Average Cost Optimality Equations Pair of coupled optimality equations J(ξ) + h(ξ) = J(ξ) = min u U E {J (φ u(ξ, Y ))}, U(ξ) def = arg min u U E {J (φ u(ξ, Y ))}, min [ḡ(ξ, u) + E {h (φ u(ξ, Y ))}], (1) u U(ξ)
20 9 Average Cost Optimality Equations Pair of coupled optimality equations J(ξ) + h(ξ) = J(ξ) = min u U E {J (φ u(ξ, Y ))}, U(ξ) def = arg min u U E {J (φ u(ξ, Y ))}, min [ḡ(ξ, u) + E {h (φ u(ξ, Y ))}], (1) u U(ξ) Constant average cost DP equation λ + h(ξ) = min u U [ḡ(ξ, u) + E {h (φ u(ξ, Y ))}]. (2)
21 9 multichain Average Cost Optimality Equations Pair of coupled optimality equations J(ξ) + h(ξ) = J(ξ) = min u U E {J (φ u(ξ, Y ))}, U(ξ) def = arg min u U E {J (φ u(ξ, Y ))}, min [ḡ(ξ, u) + E {h (φ u(ξ, Y ))}], (1) u U(ξ) Constant average cost DP equation unichain λ + h(ξ) = min u U [ḡ(ξ, u) + E {h (φ u(ξ, Y ))}]. (2)
22 9 multichain Average Cost Optimality Equations Pair of coupled optimality equations J(ξ) + h(ξ) = J(ξ) = min u U E {J (φ u(ξ, Y ))}, U(ξ) def = arg min u U E {J (φ u(ξ, Y ))}, min [ḡ(ξ, u) + E {h (φ u(ξ, Y ))}], (1) u U(ξ) Constant average cost DP equation unichain λ + h(ξ) = min u U [ḡ(ξ, u) + E {h (φ u(ξ, Y ))}]. (2) Bounded Solutions (J, h ) of (1) J ( ) = J + ( ) = J ( ) Bounded Solutions (λ, h ) of (2) J ( ) = J + ( ) = λ
23 10 Example: Non-Constant Optimal Average Cost States: {1, 2, 3, 4}, Actions: {a, b} {, 1/2 } 1 {, 1/2 } 2 {, 1 } {, 1/2 } 3 {, 1 } 4 {, 1/2 } Markov chain recurrent and aperiodic Observations: {c, d}, state 1, 3 c; state 2, 4 d Cost: g(1, a) = 1, g(1, b) = 0; g(3, a) = 0, g(3, b) = 1 ξ : ξ(1) = 1 or ξ(3) = 1, J ( ξ) = 0 ξ : ξ(1) = ξ(3) = 1/2, J (ξ) = 1/3 > 0 Non-constant optimal average cost in POMDP
24 Existence of Solutions to Constant AC DP Equation Sufficient conditions: 11 reachability and detectability [Platzman 80] interior accessibility, relative interior accessibility [Hsu, Chuang, Arapostathis 05] Constant AC: not fully understood Non-constant AC: not much has been done
25 12 Outline of the Talk Introduction to POMDP and the average cost problem Lower bounds for average cost POMDP Near-optimality of finite-state controllers
26 13 Lower Bounds for Average Cost POMDP J (ξ) J (ξ) ξ P(S). approximating processes (usually finite-state MDPs)
27 13 Lower Bounds for Average Cost POMDP J (ξ) J (ξ) ξ P(S). approximating processes (usually finite-state MDPs) Notes: J computable when spaces are finite: the bounds involve discretized approximations, finite state and control MDP algorithms
28 13 Lower Bounds for Average Cost POMDP J (ξ) J (ξ) ξ P(S). approximating processes (usually finite-state MDPs) Notes: J computable when spaces are finite: the bounds involve discretized approximations, finite state and control MDP algorithms neither J nor J have to be constant
29 13 Lower Bounds for Average Cost POMDP J (ξ) J (ξ) ξ P(S). approximating processes (usually finite-state MDPs) Notes: J computable when spaces are finite: the bounds involve discretized approximations, finite state and control MDP algorithms neither J nor J have to be constant use of lower bounds: characterize the optimal performance, provide suboptimal control
30 13 Lower Bounds for Average Cost POMDP J (ξ) J (ξ) ξ P(S). approximating processes (usually finite-state MDPs) Notes: J computable when spaces are finite: the bounds involve discretized approximations, finite state and control MDP algorithms neither J nor J have to be constant use of lower bounds: characterize the optimal performance, provide suboptimal control extension to semi-markov and constrained cases
31 14 Outline of the Analysis POMDP POMDP Fictitious Process Modified Belief MDP Modified Belief MDP Approximating POMDP
32 15 Outline of the Talk Introduction to POMDP and the average cost problem Lower bounds for average cost POMDP Near-optimality of finite-state controllers
33 16 Finite-State Controllers X X X Y 0 U0 Y 1 U1 Y2 Z Z Z internal states of controller
34 16 Finite-State Controllers X X X Y 0 U0 Y 1 U1 Y2 Z Z Z internal states of controller Key properties and advantages of FSC {(S t, Y t, Z t, U t )} jointly Markov chain, {(S t, Y t, Z t )} marginally Markov chain
35 16 Finite-State Controllers X X X Y 0 U0 Y 1 U1 Y2 Z Z Z internal states of controller Key properties and advantages of FSC {(S t, Y t, Z t, U t )} jointly Markov chain, {(S t, Y t, Z t )} marginally Markov chain asymptotic behavior well-understood; MDP theory applies
36 16 Finite-State Controllers X X X Y 0 U0 Y 1 U1 Y2 Z Z Z internal states of controller Key properties and advantages of FSC {(S t, Y t, Z t, U t )} jointly Markov chain, {(S t, Y t, Z t )} marginally Markov chain asymptotic behavior well-understood; MDP theory applies connection with piecewise linear concave approximations (Hansen 1998)
37 17 Existence of Near-Optimal FSC/Average Cost Question: Given ɛ > 0, under what conditions does there exist an ɛ-optimal FSC (with sufficiently large number of internal states)?
38 17 Existence of Near-Optimal FSC/Average Cost Question: Given ɛ > 0, under what conditions does there exist an ɛ-optimal FSC (with sufficiently large number of internal states)? Notes: We mean ɛ-optimal simultaneously for all initial beliefs (so J : flat is a necessary condition) The existence of an ɛ-optimal FSC for a given single initial belief is an open question
39 14 Existence of Near-Optimal FSC/Average Cost Question: Given ɛ > 0, under what conditions does there exist an ɛ-optimal FSC (with sufficiently large number of internal states)? Notes: We mean ɛ-optimal simultaneously for all initial beliefs (so J : flat is a necessary condition) The existence of an ɛ-optimal FSC for a given single initial belief is an open question For discounted problems, there are easy answers: There is no FSC that is simultaneously ɛ-optimal for all initial beliefs There is an ɛ-optimal FSC for a given single initial belief
40 15 Near-Optimality of Finite-State Controllers Assumption: Finite spaces, J constant Our Main Theorem: ɛ-optimal FSC. J+ = J ; and for all ɛ > 0, there exists an
41 15 Near-Optimality of Finite-State Controllers Assumption: Finite spaces, J constant Our Main Theorem: ɛ-optimal FSC. J+ = J ; and for all ɛ > 0, there exists an Recall definitions: π = {µ 0, µ 1,..., }, µ 0 ( ), µ t (h t, ) P(U) Jk π (ξ) = EPξ,π J π (ξ) = lim inf k 1 k Jπ k (ξ) J (ξ) = inf π Π Jπ (ξ) k 1 g(s t, U t ) t=0
42 15 Near-Optimality of Finite-State Controllers Assumption: Finite spaces, J constant Our Main Theorem: ɛ-optimal FSC. J+ = J ; and for all ɛ > 0, there exists an Recall definitions: π = {µ 0, µ 1,..., }, µ 0 ( ), µ t (h t, ) P(U) Jk π (ξ) = EPξ,π Lemma: J π (ξ) = lim inf k 1 k Jπ k (ξ) J (ξ) = inf π Π Jπ (ξ) Jk π(ξ) linear; Jπ (ξ) concave; J (ξ) concave. k 1 g(s t, U t ) t=0
43 15 Near-Optimality of Finite-State Controllers Assumption: Finite spaces, J constant Our Main Theorem: ɛ-optimal FSC. J+ = J ; and for all ɛ > 0, there exists an Recall definitions: π = {µ 0, µ 1,..., }, µ 0 ( ), µ t (h t, ) P(U) Jk π (ξ) = EPξ,π Lemma: J π (ξ) = lim inf k 1 k Jπ k (ξ) J (ξ) = inf π Π Jπ (ξ) Jk π(ξ) linear; Jπ (ξ) concave; J (ξ) concave. k 1 g(s t, U t ) t=0 1st Key Observation: Suppose J cannot happen... J π constant. Then, the first case J π J J
44 15 Near-Optimality of Finite-State Controllers Assumption: Finite spaces, J constant Our Main Theorem: ɛ-optimal FSC. J+ = J ; and for all ɛ > 0, there exists an Recall definitions: π = {µ 0, µ 1,..., }, µ 0 ( ), µ t (h t, ) P(U) Jk π (ξ) = EPξ,π Lemma: J π (ξ) = lim inf k 1 k Jπ k (ξ) J (ξ) = inf π Π Jπ (ξ) Jk π(ξ) linear; Jπ (ξ) concave; J (ξ) concave. k 1 g(s t, U t ) t=0 1st Key Observation: Suppose J cannot happen... J π constant. Then, the first case J π J J There must exist π nearly optimal for some interior ξ, with J π almost flat
45 First Result on Independence from Initial Belief Proposition: For all ɛ > 0, there exists an ɛ-liminf optimal policy π that does not functionally depend on the initial belief ξ, i.e., 16 J π (ξ) J + ɛ, ξ P(S).
46 First Result on Independence from Initial Belief Proposition: For all ɛ > 0, there exists an ɛ-liminf optimal policy π that does not functionally depend on the initial belief ξ, i.e., 16 J π (ξ) J + ɛ, ξ P(S). 2nd Key Observation: There must exist k 0, with 1 k 0 J π k 0 uniformly close to J π 1 k 1 J π k 1 1 k 2 J π k 2 J π = lim inf k 1 k Jπ k
47 17 Finite-Stage Uniform Closeness Lemma Lemma: For all ɛ > 0, there exists π 0 Π and an integer k 0 (depending on π 0 ) such that 1 k 0 J π 0 k 0 (ξ) J ɛ, ξ P(S).
48 17 Finite-Stage Uniform Closeness Lemma Lemma: For all ɛ > 0, there exists π 0 Π and an integer k 0 (depending on π 0 ) such that 1 k 0 J π 0 k 0 (ξ) J ɛ, ξ P(S). Final construction: Construct infinite-stage policy by replication of the finite-stage policy: Form a new policy π 1 that has a finite-length history window : π 0 = {µ 0, µ 1,..., µ k0 1,...} π 1 = {µ 0, µ 1,..., µ k0 1, µ k 0, µ k 0 +1,..., µ 2k 0 1,...} µ t (h t, ) = µ k(t) (δ k(t) (h t ), ), k(t) def = mod(t, k 0 ), where δ k(t) extracts the last length- k(t) segment of a length-t history h t.
49 17 Finite-Stage Uniform Closeness Lemma Lemma: For all ɛ > 0, there exists π 0 Π and an integer k 0 (depending on π 0 ) such that 1 k 0 J π 0 k 0 (ξ) J ɛ, ξ P(S). Final construction: Construct infinite-stage policy by replication of the finite-stage policy: Form a new policy π 1 that has a finite-length history window : π 0 = {µ 0, µ 1,..., µ k0 1,...} π 1 = {µ 0, µ 1,..., µ k0 1, µ k 0, µ k 0 +1,..., µ 2k 0 1,...} µ t (h t, ) = µ k(t) (δ k(t) (h t ), ), k(t) def = mod(t, k 0 ), where δ k(t) extracts the last length- k(t) segment of a length-t history h t. J π 0 ( ): uniformly close to J = Jπ 1 ( ): uniformly close to J
50 17 Finite-Stage Uniform Closeness Lemma Lemma: For all ɛ > 0, there exists π 0 Π and an integer k 0 (depending on π 0 ) such that 1 k 0 J π 0 k 0 (ξ) J ɛ, ξ P(S). Final construction: Construct infinite-stage policy by replication of the finite-stage policy: Form a new policy π 1 that has a finite-length history window : π 0 = {µ 0, µ 1,..., µ k0 1,...} π 1 = {µ 0, µ 1,..., µ k0 1, µ k 0, µ k 0 +1,..., µ 2k 0 1,...} µ t (h t, ) = µ k(t) (δ k(t) (h t ), ), k(t) def = mod(t, k 0 ), where δ k(t) extracts the last length- k(t) segment of a length-t history h t. J π 0 ( ): uniformly close to J = Jπ 1 ( ): uniformly close to J For finite spaces: π 1 is FSC = J π 1 (ξ) = Jπ 1 + (ξ), ξ P(S).
51 18 Near-Optimality of FSC Assumption: Finite spaces, J constant Theorem: J + = J for all ɛ > 0, there is a FSC that is ɛ-optimal, and does not functionally depend on the initial belief ξ. Note: The ɛ-optimal FSC is also a finite-history controller Intuitive implication: If J does not depend on the initial belief, old information becomes increasingly obsolete
Optimistic Policy Iteration and Q-learning in Dynamic Programming
Optimistic Policy Iteration and Q-learning in Dynamic Programming Dimitri P. Bertsekas Laboratory for Information and Decision Systems Massachusetts Institute of Technology November 2010 INFORMS, Austin,
More informationInfinite-Horizon Average Reward Markov Decision Processes
Infinite-Horizon Average Reward Markov Decision Processes Dan Zhang Leeds School of Business University of Colorado at Boulder Dan Zhang, Spring 2012 Infinite Horizon Average Reward MDP 1 Outline The average
More informationComputational complexity estimates for value and policy iteration algorithms for total-cost and average-cost Markov decision processes
Computational complexity estimates for value and policy iteration algorithms for total-cost and average-cost Markov decision processes Jefferson Huang Dept. Applied Mathematics and Statistics Stony Brook
More informationA Function Approximation Approach to Estimation of Policy Gradient for POMDP with Structured Policies
A Function Approximation Approach to Estimation of Policy Gradient for POMDP with Structured Policies Huizhen Yu Lab for Information and Decision Systems Massachusetts Institute of Technology Cambridge,
More informationReductions Of Undiscounted Markov Decision Processes and Stochastic Games To Discounted Ones. Jefferson Huang
Reductions Of Undiscounted Markov Decision Processes and Stochastic Games To Discounted Ones Jefferson Huang School of Operations Research and Information Engineering Cornell University November 16, 2016
More informationChristopher Watkins and Peter Dayan. Noga Zaslavsky. The Hebrew University of Jerusalem Advanced Seminar in Deep Learning (67679) November 1, 2015
Q-Learning Christopher Watkins and Peter Dayan Noga Zaslavsky The Hebrew University of Jerusalem Advanced Seminar in Deep Learning (67679) November 1, 2015 Noga Zaslavsky Q-Learning (Watkins & Dayan, 1992)
More informationPartially Observable Markov Decision Processes (POMDPs)
Partially Observable Markov Decision Processes (POMDPs) Geoff Hollinger Sequential Decision Making in Robotics Spring, 2011 *Some media from Reid Simmons, Trey Smith, Tony Cassandra, Michael Littman, and
More informationCentral-limit approach to risk-aware Markov decision processes
Central-limit approach to risk-aware Markov decision processes Jia Yuan Yu Concordia University November 27, 2015 Joint work with Pengqian Yu and Huan Xu. Inventory Management 1 1 Look at current inventory
More informationAn Empirical Algorithm for Relative Value Iteration for Average-cost MDPs
2015 IEEE 54th Annual Conference on Decision and Control CDC December 15-18, 2015. Osaka, Japan An Empirical Algorithm for Relative Value Iteration for Average-cost MDPs Abhishek Gupta Rahul Jain Peter
More informationValue and Policy Iteration
Chapter 7 Value and Policy Iteration 1 For infinite horizon problems, we need to replace our basic computational tool, the DP algorithm, which we used to compute the optimal cost and policy for finite
More informationStochastic Dynamic Programming. Jesus Fernandez-Villaverde University of Pennsylvania
Stochastic Dynamic Programming Jesus Fernande-Villaverde University of Pennsylvania 1 Introducing Uncertainty in Dynamic Programming Stochastic dynamic programming presents a very exible framework to handle
More informationHidden Markov Models (HMM) and Support Vector Machine (SVM)
Hidden Markov Models (HMM) and Support Vector Machine (SVM) Professor Joongheon Kim School of Computer Science and Engineering, Chung-Ang University, Seoul, Republic of Korea 1 Hidden Markov Models (HMM)
More information10 Robotic Exploration and Information Gathering
NAVARCH/EECS 568, ROB 530 - Winter 2018 10 Robotic Exploration and Information Gathering Maani Ghaffari April 2, 2018 Robotic Information Gathering: Exploration and Monitoring In information gathering
More informationWeighted Sup-Norm Contractions in Dynamic Programming: A Review and Some New Applications
May 2012 Report LIDS - 2884 Weighted Sup-Norm Contractions in Dynamic Programming: A Review and Some New Applications Dimitri P. Bertsekas Abstract We consider a class of generalized dynamic programming
More informationLecture notes for Analysis of Algorithms : Markov decision processes
Lecture notes for Analysis of Algorithms : Markov decision processes Lecturer: Thomas Dueholm Hansen June 6, 013 Abstract We give an introduction to infinite-horizon Markov decision processes (MDPs) with
More informationMarkov decision processes and interval Markov chains: exploiting the connection
Markov decision processes and interval Markov chains: exploiting the connection Mingmei Teo Supervisors: Prof. Nigel Bean, Dr Joshua Ross University of Adelaide July 10, 2013 Intervals and interval arithmetic
More informationMarkov decision processes
CS 2740 Knowledge representation Lecture 24 Markov decision processes Milos Hauskrecht milos@cs.pitt.edu 5329 Sennott Square Administrative announcements Final exam: Monday, December 8, 2008 In-class Only
More informationBayesian reinforcement learning and partially observable Markov decision processes November 6, / 24
and partially observable Markov decision processes Christos Dimitrakakis EPFL November 6, 2013 Bayesian reinforcement learning and partially observable Markov decision processes November 6, 2013 1 / 24
More informationChapter 2 SOME ANALYTICAL TOOLS USED IN THE THESIS
Chapter 2 SOME ANALYTICAL TOOLS USED IN THE THESIS 63 2.1 Introduction In this chapter we describe the analytical tools used in this thesis. They are Markov Decision Processes(MDP), Markov Renewal process
More informationMarkov Decision Processes
Markov Decision Processes Lecture notes for the course Games on Graphs B. Srivathsan Chennai Mathematical Institute, India 1 Markov Chains We will define Markov chains in a manner that will be useful to
More informationMARKOV DECISION PROCESSES (MDP) AND REINFORCEMENT LEARNING (RL) Versione originale delle slide fornita dal Prof. Francesco Lo Presti
1 MARKOV DECISION PROCESSES (MDP) AND REINFORCEMENT LEARNING (RL) Versione originale delle slide fornita dal Prof. Francesco Lo Presti Historical background 2 Original motivation: animal learning Early
More informationAn Adaptive Clustering Method for Model-free Reinforcement Learning
An Adaptive Clustering Method for Model-free Reinforcement Learning Andreas Matt and Georg Regensburger Institute of Mathematics University of Innsbruck, Austria {andreas.matt, georg.regensburger}@uibk.ac.at
More informationThe convergence limit of the temporal difference learning
The convergence limit of the temporal difference learning Ryosuke Nomura the University of Tokyo September 3, 2013 1 Outline Reinforcement Learning Convergence limit Construction of the feature vector
More informationDecision Theory: Markov Decision Processes
Decision Theory: Markov Decision Processes CPSC 322 Lecture 33 March 31, 2006 Textbook 12.5 Decision Theory: Markov Decision Processes CPSC 322 Lecture 33, Slide 1 Lecture Overview Recap Rewards and Policies
More informationReinforcement Learning
Reinforcement Learning March May, 2013 Schedule Update Introduction 03/13/2015 (10:15-12:15) Sala conferenze MDPs 03/18/2015 (10:15-12:15) Sala conferenze Solving MDPs 03/20/2015 (10:15-12:15) Aula Alpha
More informationOn the reduction of total cost and average cost MDPs to discounted MDPs
On the reduction of total cost and average cost MDPs to discounted MDPs Jefferson Huang School of Operations Research and Information Engineering Cornell University July 12, 2017 INFORMS Applied Probability
More informationWE consider finite-state Markov decision processes
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 54, NO. 7, JULY 2009 1515 Convergence Results for Some Temporal Difference Methods Based on Least Squares Huizhen Yu and Dimitri P. Bertsekas Abstract We consider
More informationRobust Markov Decision Processes
Robust Markov Decision Processes Wolfram Wiesemann, Daniel Kuhn and Berç Rustem February 9, 2012 Abstract Markov decision processes (MDPs) are powerful tools for decision making in uncertain dynamic environments.
More informationApproximate dynamic programming for stochastic reachability
Approximate dynamic programming for stochastic reachability Nikolaos Kariotoglou, Sean Summers, Tyler Summers, Maryam Kamgarpour and John Lygeros Abstract In this work we illustrate how approximate dynamic
More informationPOMDPs and Policy Gradients
POMDPs and Policy Gradients MLSS 2006, Canberra Douglas Aberdeen Canberra Node, RSISE Building Australian National University 15th February 2006 Outline 1 Introduction What is Reinforcement Learning? Types
More informationOptimal Control of Partially Observable Piecewise Deterministic Markov Processes
Optimal Control of Partially Observable Piecewise Deterministic Markov Processes Nicole Bäuerle based on a joint work with D. Lange Wien, April 2018 Outline Motivation PDMP with Partial Observation Controlled
More informationRegular Policies in Abstract Dynamic Programming
August 2016 (Revised January 2017) Report LIDS-P-3173 Regular Policies in Abstract Dynamic Programming Dimitri P. Bertsekas Abstract We consider challenging dynamic programming models where the associated
More informationTheory of Stochastic Processes 3. Generating functions and their applications
Theory of Stochastic Processes 3. Generating functions and their applications Tomonari Sei sei@mist.i.u-tokyo.ac.jp Department of Mathematical Informatics, University of Tokyo April 20, 2017 http://www.stat.t.u-tokyo.ac.jp/~sei/lec.html
More informationProcedia Computer Science 00 (2011) 000 6
Procedia Computer Science (211) 6 Procedia Computer Science Complex Adaptive Systems, Volume 1 Cihan H. Dagli, Editor in Chief Conference Organized by Missouri University of Science and Technology 211-
More information6.231 DYNAMIC PROGRAMMING LECTURE 7 LECTURE OUTLINE
6.231 DYNAMIC PROGRAMMING LECTURE 7 LECTURE OUTLINE DP for imperfect state info Sufficient statistics Conditional state distribution as a sufficient statistic Finite-state systems Examples 1 REVIEW: IMPERFECT
More informationEfficient Maximization in Solving POMDPs
Efficient Maximization in Solving POMDPs Zhengzhu Feng Computer Science Department University of Massachusetts Amherst, MA 01003 fengzz@cs.umass.edu Shlomo Zilberstein Computer Science Department University
More informationA Review of the E 3 Algorithm: Near-Optimal Reinforcement Learning in Polynomial Time
A Review of the E 3 Algorithm: Near-Optimal Reinforcement Learning in Polynomial Time April 16, 2016 Abstract In this exposition we study the E 3 algorithm proposed by Kearns and Singh for reinforcement
More informationON THE POLICY IMPROVEMENT ALGORITHM IN CONTINUOUS TIME
ON THE POLICY IMPROVEMENT ALGORITHM IN CONTINUOUS TIME SAUL D. JACKA AND ALEKSANDAR MIJATOVIĆ Abstract. We develop a general approach to the Policy Improvement Algorithm (PIA) for stochastic control problems
More informationSemi-Infinite Relaxations for a Dynamic Knapsack Problem
Semi-Infinite Relaxations for a Dynamic Knapsack Problem Alejandro Toriello joint with Daniel Blado, Weihong Hu Stewart School of Industrial and Systems Engineering Georgia Institute of Technology MIT
More informationLecture 3: Markov Decision Processes
Lecture 3: Markov Decision Processes Joseph Modayil 1 Markov Processes 2 Markov Reward Processes 3 Markov Decision Processes 4 Extensions to MDPs Markov Processes Introduction Introduction to MDPs Markov
More information1 Problem Formulation
Book Review Self-Learning Control of Finite Markov Chains by A. S. Poznyak, K. Najim, and E. Gómez-Ramírez Review by Benjamin Van Roy This book presents a collection of work on algorithms for learning
More informationUniversity of Warwick, EC9A0 Maths for Economists Lecture Notes 10: Dynamic Programming
University of Warwick, EC9A0 Maths for Economists 1 of 63 University of Warwick, EC9A0 Maths for Economists Lecture Notes 10: Dynamic Programming Peter J. Hammond Autumn 2013, revised 2014 University of
More informationArtificial Intelligence & Sequential Decision Problems
Artificial Intelligence & Sequential Decision Problems (CIV6540 - Machine Learning for Civil Engineers) Professor: James-A. Goulet Département des génies civil, géologique et des mines Chapter 15 Goulet
More informationMonte Carlo Linear Algebra: A Review and Recent Results
Monte Carlo Linear Algebra: A Review and Recent Results Dimitri P. Bertsekas Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology Caradache, France, 2012 *Athena
More informationLet (Ω, F) be a measureable space. A filtration in discrete time is a sequence of. F s F t
2.2 Filtrations Let (Ω, F) be a measureable space. A filtration in discrete time is a sequence of σ algebras {F t } such that F t F and F t F t+1 for all t = 0, 1,.... In continuous time, the second condition
More informationDecision Theory: Q-Learning
Decision Theory: Q-Learning CPSC 322 Decision Theory 5 Textbook 12.5 Decision Theory: Q-Learning CPSC 322 Decision Theory 5, Slide 1 Lecture Overview 1 Recap 2 Asynchronous Value Iteration 3 Q-Learning
More informationAbstract Dynamic Programming
Abstract Dynamic Programming Dimitri P. Bertsekas Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology Overview of the Research Monograph Abstract Dynamic Programming"
More informationConsistency of the maximum likelihood estimator for general hidden Markov models
Consistency of the maximum likelihood estimator for general hidden Markov models Jimmy Olsson Centre for Mathematical Sciences Lund University Nordstat 2012 Umeå, Sweden Collaborators Hidden Markov models
More informationMDP Preliminaries. Nan Jiang. February 10, 2019
MDP Preliminaries Nan Jiang February 10, 2019 1 Markov Decision Processes In reinforcement learning, the interactions between the agent and the environment are often described by a Markov Decision Process
More informationOn the existence of stationary optimal policies for partially observed MDPs under the long-run average cost criterion
Systems & Control Letters 55 (2006) 165 173 www.elsevier.com/locate/sysconle On the existence of stationary optimal policies for partially observed MDPs under the long-run average cost criterion Shun-Pin
More informationMarkov Decision Processes and Solving Finite Problems. February 8, 2017
Markov Decision Processes and Solving Finite Problems February 8, 2017 Overview of Upcoming Lectures Feb 8: Markov decision processes, value iteration, policy iteration Feb 13: Policy gradients Feb 15:
More informationAlternative Characterization of Ergodicity for Doubly Stochastic Chains
Alternative Characterization of Ergodicity for Doubly Stochastic Chains Behrouz Touri and Angelia Nedić Abstract In this paper we discuss the ergodicity of stochastic and doubly stochastic chains. We define
More informationReinforcement Learning and Optimal Control. ASU, CSE 691, Winter 2019
Reinforcement Learning and Optimal Control ASU, CSE 691, Winter 2019 Dimitri P. Bertsekas dimitrib@mit.edu Lecture 8 Bertsekas Reinforcement Learning 1 / 21 Outline 1 Review of Infinite Horizon Problems
More informationOn Prediction and Planning in Partially Observable Markov Decision Processes with Large Observation Sets
On Prediction and Planning in Partially Observable Markov Decision Processes with Large Observation Sets Pablo Samuel Castro pcastr@cs.mcgill.ca McGill University Joint work with: Doina Precup and Prakash
More informationBayesian Congestion Control over a Markovian Network Bandwidth Process: A multiperiod Newsvendor Problem
Bayesian Congestion Control over a Markovian Network Bandwidth Process: A multiperiod Newsvendor Problem Parisa Mansourifard 1/37 Bayesian Congestion Control over a Markovian Network Bandwidth Process:
More informationPlanning and Acting in Partially Observable Stochastic Domains
Planning and Acting in Partially Observable Stochastic Domains Leslie Pack Kaelbling*, Michael L. Littman**, Anthony R. Cassandra*** *Computer Science Department, Brown University, Providence, RI, USA
More informationInfinite-Horizon Discounted Markov Decision Processes
Infinite-Horizon Discounted Markov Decision Processes Dan Zhang Leeds School of Business University of Colorado at Boulder Dan Zhang, Spring 2012 Infinite Horizon Discounted MDP 1 Outline The expected
More informationA Geometric Framework for Nonconvex Optimization Duality using Augmented Lagrangian Functions
A Geometric Framework for Nonconvex Optimization Duality using Augmented Lagrangian Functions Angelia Nedić and Asuman Ozdaglar April 15, 2006 Abstract We provide a unifying geometric framework for the
More informationAverage Reward Parameters
Simulation-Based Optimization of Markov Reward Processes: Implementation Issues Peter Marbach 2 John N. Tsitsiklis 3 Abstract We consider discrete time, nite state space Markov reward processes which depend
More informationStochastic Primal-Dual Methods for Reinforcement Learning
Stochastic Primal-Dual Methods for Reinforcement Learning Alireza Askarian 1 Amber Srivastava 1 1 Department of Mechanical Engineering University of Illinois at Urbana Champaign Big Data Optimization,
More informationOutline. CSE 573: Artificial Intelligence Autumn Agent. Partial Observability. Markov Decision Process (MDP) 10/31/2012
CSE 573: Artificial Intelligence Autumn 2012 Reasoning about Uncertainty & Hidden Markov Models Daniel Weld Many slides adapted from Dan Klein, Stuart Russell, Andrew Moore & Luke Zettlemoyer 1 Outline
More informationBayesian Congestion Control over a Markovian Network Bandwidth Process
Bayesian Congestion Control over a Markovian Network Bandwidth Process Parisa Mansourifard 1/30 Bayesian Congestion Control over a Markovian Network Bandwidth Process Parisa Mansourifard (USC) Joint work
More informationMean Field Competitive Binary MDPs and Structured Solutions
Mean Field Competitive Binary MDPs and Structured Solutions Minyi Huang School of Mathematics and Statistics Carleton University Ottawa, Canada MFG217, UCLA, Aug 28 Sept 1, 217 1 / 32 Outline of talk The
More informationComplexity of stochastic branch and bound methods for belief tree search in Bayesian reinforcement learning
Complexity of stochastic branch and bound methods for belief tree search in Bayesian reinforcement learning Christos Dimitrakakis Informatics Institute, University of Amsterdam, Amsterdam, The Netherlands
More informationMarkov Decision Processes and Dynamic Programming
Markov Decision Processes and Dynamic Programming A. LAZARIC (SequeL Team @INRIA-Lille) Ecole Centrale - Option DAD SequeL INRIA Lille EC-RL Course In This Lecture A. LAZARIC Markov Decision Processes
More informationREMARKS ON THE EXISTENCE OF SOLUTIONS IN MARKOV DECISION PROCESSES. Emmanuel Fernández-Gaucherand, Aristotle Arapostathis, and Steven I.
REMARKS ON THE EXISTENCE OF SOLUTIONS TO THE AVERAGE COST OPTIMALITY EQUATION IN MARKOV DECISION PROCESSES Emmanuel Fernández-Gaucherand, Aristotle Arapostathis, and Steven I. Marcus Department of Electrical
More informationIntroduction to Empirical Processes and Semiparametric Inference Lecture 08: Stochastic Convergence
Introduction to Empirical Processes and Semiparametric Inference Lecture 08: Stochastic Convergence Michael R. Kosorok, Ph.D. Professor and Chair of Biostatistics Professor of Statistics and Operations
More informationBias-Variance Error Bounds for Temporal Difference Updates
Bias-Variance Bounds for Temporal Difference Updates Michael Kearns AT&T Labs mkearns@research.att.com Satinder Singh AT&T Labs baveja@research.att.com Abstract We give the first rigorous upper bounds
More informationMATH4406 Assignment 5
MATH4406 Assignment 5 Patrick Laub (ID: 42051392) October 7, 2014 1 The machine replacement model 1.1 Real-world motivation Consider the machine to be the entire world. Over time the creator has running
More informationPrioritized Sweeping Converges to the Optimal Value Function
Technical Report DCS-TR-631 Prioritized Sweeping Converges to the Optimal Value Function Lihong Li and Michael L. Littman {lihong,mlittman}@cs.rutgers.edu RL 3 Laboratory Department of Computer Science
More informationOn the Convergence of Optimistic Policy Iteration
Journal of Machine Learning Research 3 (2002) 59 72 Submitted 10/01; Published 7/02 On the Convergence of Optimistic Policy Iteration John N. Tsitsiklis LIDS, Room 35-209 Massachusetts Institute of Technology
More informationInformation-Theoretic Privacy for Smart Metering Systems with a Rechargeable Battery
1 Information-Theoretic Privacy for Smart Metering Systems with a Rechargeable Battery Simon Li, Ashish Khisti, and Aditya Mahajan Abstract Smart-metering systems report electricity usage of a user to
More informationRobustness of policies in Constrained Markov Decision Processes
1 Robustness of policies in Constrained Markov Decision Processes Alexander Zadorojniy and Adam Shwartz, Senior Member, IEEE Abstract We consider the optimization of finite-state, finite-action Markov
More informationJournal of Computational and Applied Mathematics. Projected equation methods for approximate solution of large linear systems
Journal of Computational and Applied Mathematics 227 2009) 27 50 Contents lists available at ScienceDirect Journal of Computational and Applied Mathematics journal homepage: www.elsevier.com/locate/cam
More informationBasics of reinforcement learning
Basics of reinforcement learning Lucian Buşoniu TMLSS, 20 July 2018 Main idea of reinforcement learning (RL) Learn a sequential decision policy to optimize the cumulative performance of an unknown system
More informationC.7. Numerical series. Pag. 147 Proof of the converging criteria for series. Theorem 5.29 (Comparison test) Let a k and b k be positive-term series
C.7 Numerical series Pag. 147 Proof of the converging criteria for series Theorem 5.29 (Comparison test) Let and be positive-term series such that 0, for any k 0. i) If the series converges, then also
More informationReinforcement Learning as Variational Inference: Two Recent Approaches
Reinforcement Learning as Variational Inference: Two Recent Approaches Rohith Kuditipudi Duke University 11 August 2017 Outline 1 Background 2 Stein Variational Policy Gradient 3 Soft Q-Learning 4 Closing
More informationTowards Uncertainty-Aware Path Planning On Road Networks Using Augmented-MDPs. Lorenzo Nardi and Cyrill Stachniss
Towards Uncertainty-Aware Path Planning On Road Networks Using Augmented-MDPs Lorenzo Nardi and Cyrill Stachniss Navigation under uncertainty C B C B A A 2 `B` is the most likely position C B C B A A 3
More informationLECTURE 4 LECTURE OUTLINE
LECTURE 4 LECTURE OUTLINE Relative interior and closure Algebra of relative interiors and closures Continuity of convex functions Closures of functions Reading: Section 1.3 All figures are courtesy of
More informationGeneral Theory of Large Deviations
Chapter 30 General Theory of Large Deviations A family of random variables follows the large deviations principle if the probability of the variables falling into bad sets, representing large deviations
More informationAn Introduction to Markov Decision Processes. MDP Tutorial - 1
An Introduction to Markov Decision Processes Bob Givan Purdue University Ron Parr Duke University MDP Tutorial - 1 Outline Markov Decision Processes defined (Bob) Objective functions Policies Finding Optimal
More informationCAP Plan, Activity, and Intent Recognition
CAP6938-02 Plan, Activity, and Intent Recognition Lecture 10: Sequential Decision-Making Under Uncertainty (part 1) MDPs and POMDPs Instructor: Dr. Gita Sukthankar Email: gitars@eecs.ucf.edu SP2-1 Reminder
More informationMATH 722, COMPLEX ANALYSIS, SPRING 2009 PART 5
MATH 722, COMPLEX ANALYSIS, SPRING 2009 PART 5.. The Arzela-Ascoli Theorem.. The Riemann mapping theorem Let X be a metric space, and let F be a family of continuous complex-valued functions on X. We have
More informationToday s Outline. Recap: MDPs. Bellman Equations. Q-Value Iteration. Bellman Backup 5/7/2012. CSE 473: Artificial Intelligence Reinforcement Learning
CSE 473: Artificial Intelligence Reinforcement Learning Dan Weld Today s Outline Reinforcement Learning Q-value iteration Q-learning Exploration / exploitation Linear function approximation Many slides
More informationControlling probabilistic systems under partial observation an automata and verification perspective
Controlling probabilistic systems under partial observation an automata and verification perspective Nathalie Bertrand, Inria Rennes, France Uncertainty in Computation Workshop October 4th 2016, Simons
More informationLearning in Zero-Sum Team Markov Games using Factored Value Functions
Learning in Zero-Sum Team Markov Games using Factored Value Functions Michail G. Lagoudakis Department of Computer Science Duke University Durham, NC 27708 mgl@cs.duke.edu Ronald Parr Department of Computer
More informationFinding the Value of Information About a State Variable in a Markov Decision Process 1
05/25/04 1 Finding the Value of Information About a State Variable in a Markov Decision Process 1 Gilvan C. Souza The Robert H. Smith School of usiness, The University of Maryland, College Park, MD, 20742
More informationA projection algorithm for strictly monotone linear complementarity problems.
A projection algorithm for strictly monotone linear complementarity problems. Erik Zawadzki Department of Computer Science epz@cs.cmu.edu Geoffrey J. Gordon Machine Learning Department ggordon@cs.cmu.edu
More informationReinforcement Learning as Classification Leveraging Modern Classifiers
Reinforcement Learning as Classification Leveraging Modern Classifiers Michail G. Lagoudakis and Ronald Parr Department of Computer Science Duke University Durham, NC 27708 Machine Learning Reductions
More informationInformation theoretic perspectives on learning algorithms
Information theoretic perspectives on learning algorithms Varun Jog University of Wisconsin - Madison Departments of ECE and Mathematics Shannon Channel Hangout! May 8, 2018 Jointly with Adrian Tovar-Lopez
More informationExpectation Propagation in Factor Graphs: A Tutorial
DRAFT: Version 0.1, 28 October 2005. Do not distribute. Expectation Propagation in Factor Graphs: A Tutorial Charles Sutton October 28, 2005 Abstract Expectation propagation is an important variational
More informationStatistical Models and Algorithms for Real-Time Anomaly Detection Using Multi-Modal Data
Statistical Models and Algorithms for Real-Time Anomaly Detection Using Multi-Modal Data Taposh Banerjee University of Texas at San Antonio Joint work with Gene Whipps (US Army Research Laboratory) Prudhvi
More informationGrundlagen der Künstlichen Intelligenz
Grundlagen der Künstlichen Intelligenz Formal models of interaction Daniel Hennes 27.11.2017 (WS 2017/18) University Stuttgart - IPVS - Machine Learning & Robotics 1 Today Taxonomy of domains Models of
More informationBiasing Approximate Dynamic Programming with a Lower Discount Factor
Biasing Approximate Dynamic Programming with a Lower Discount Factor Marek Petrik, Bruno Scherrer To cite this version: Marek Petrik, Bruno Scherrer. Biasing Approximate Dynamic Programming with a Lower
More information1 MDP Value Iteration Algorithm
CS 0. - Active Learning Problem Set Handed out: 4 Jan 009 Due: 9 Jan 009 MDP Value Iteration Algorithm. Implement the value iteration algorithm given in the lecture. That is, solve Bellman s equation using
More informationOptimism in the Face of Uncertainty Should be Refutable
Optimism in the Face of Uncertainty Should be Refutable Ronald ORTNER Montanuniversität Leoben Department Mathematik und Informationstechnolgie Franz-Josef-Strasse 18, 8700 Leoben, Austria, Phone number:
More informationQ-Learning for Markov Decision Processes*
McGill University ECSE 506: Term Project Q-Learning for Markov Decision Processes* Authors: Khoa Phan khoa.phan@mail.mcgill.ca Sandeep Manjanna sandeep.manjanna@mail.mcgill.ca (*Based on: Convergence of
More informationOnline Learning for Markov Decision Processes Applied to Multi-Agent Systems
Online Learning for Markov Decision Processes Applied to Multi-Agent Systems Mahmoud El Chamie Behçet Açıkmeşe Mehran Mesbahi Abstract Online learning is the process of providing online control decisions
More informationMATH 56A: STOCHASTIC PROCESSES CHAPTER 2
MATH 56A: STOCHASTIC PROCESSES CHAPTER 2 2. Countable Markov Chains I started Chapter 2 which talks about Markov chains with a countably infinite number of states. I did my favorite example which is on
More information, and rewards and transition matrices as shown below:
CSE 50a. Assignment 7 Out: Tue Nov Due: Thu Dec Reading: Sutton & Barto, Chapters -. 7. Policy improvement Consider the Markov decision process (MDP) with two states s {0, }, two actions a {0, }, discount
More information