Dynamic Programming and Optimal Control Volume 1 SECOND EDITION

By Dimitri P. Bertsekas. 2000

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Trajectory-wise Control Variates for Variance Reduction in Policy Gradient Methods (2019, background): “We show that when dynamics is the MDP is unknown, an ordering is feasible to implement, if and only if,Rk appears before Sk+1..h for all t ≤ k h.”
DeepMind Control Suite (2018, background): “Termination and Discount: Control problems are classified as finite-horizon, firstexit and infinite-horizon (Bertsekas, 1995).”
Noisy Networks for Exploration (2017, background): “1 Markov Decision Processes and Reinforcement Learning MDPs model stochastic, discrete-time and finite action-state space control problems [5, 6, 28].”
Fast Marching Tree: a Fast Marching Sampling-Based Method for Optimal Motion Planning in Many Dimensions (2013, methodology): “In the interest of brevity, we will only discuss the required modifications to existing theorems, rather than proving everything from scratch.”
Offline Meta Reinforcement Learning (2020, background): “A task Ti is described by a Markov Decision Process (MDP [2])Mi = (S,A,Ri,Pi), where the state space S and the action space A are shared across tasks, and Ri and Pi are task specific reward and transition functions.”
Markov Decision Processes with Time-varying Transition Probabilities and Rewards (2019, background): “Moreover, it is proved in (Bertsekas, 2005) that the total reward h̃(r, g,P, π) is bounded by the following:”
Randomized Linear Programming Solves the Discounted Markov Decision Problem In Nearly-Linear Running Time (2017, methodology): “For more detailed surveys on MDP and its solution methods, we refer the readers to the textbooks [3, 5, 24, 4] and references therein.”
Using deep Q-learning to understand the tax evasion behavior of risk-averse firms (2018, background): “That assumption kept the firm’s state and decision spaces conveniently small (it implied, for example, that the firm’s optimal decision is to either be completely honest or to conceal as much profit as possible, eliminating “intermediate” options), making the problem of optimizing taxpayer behavior ”
Randomized Linear Programming Solves the Discounted Markov Decision Problem In Nearly-Linear (Sometimes Sublinear) Running Time (2017, methodology): “According to the theory of dynamic programming [26, 3], a vector v∗ is the optimal value function to the DMDP if and only if it satisfies the following |S| × |S| system of equations, known as the Bellman equation, given by”
dm_control: Software and Tasks for Continuous Control (2020, background): “Termination and Discount: Control problems are usually classified as finite-horizon, first-exit and infinite-horizon (Bertsekas, 1995).”