Dynamic Programming and Optimal Control Problem Set: Deterministic Systems and the Shortest Path Problem

By R. Weber. 2008

How researchers see this paper

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

In Algorithms & Complexity

Citations in the corpus

OpenAI Gym (2016, background): “RL has a rich mathematical theory and has found a variety of practical applications [1].”
Taming the Noise in Reinforcement Learning via Soft Updates (2015, methodology): “Adding the penalty term (10) to the cost function (1) gives called the free-energy function by analogy with a similar quantity in statistical mechanics [10].”
Learning and Querying Fast Generative Models for Reinforcement Learning (2018, background): “, 2016a), dynamic programming (Bertsekas et al., 1995) or generating synthetic experiences (Sutton, 1991).”
Double Reinforcement Learning for Efficient Off-Policy Evaluation in Markov Decision Processes (2019, methodology): “To do this, we will leverage a recursive deﬁnition of the q -functions (Bertsekas, 2012).”
Reverse Curriculum Generation for Reinforcement Learning (2017, methodology): “Reinforcement learning tackles this problem with methods like policy search, in which trajectories are iteratively collected and used to improve the current policy [28, 29, 30].”
A Stochastic Optimal Control Approach for Power Management in Plug-In Hybrid Electric Vehicles (2011, background): “Stochastic dynamic programming (SDP) methods are particularly appealing in this context, despite their well-recognized computational complexity [20], because of their ability to optimize a power split map for a probabilistic distribution of many drive cycles, rather than a single cycle [21] [22]…”
Residential Demand Response Using Reinforcement Learning (2010, background): “Bellman’s equation [18] describes the optimality condition for a discounted cost MDP.”
Distance Transforms of Sampled Functions (2012, methodology): “For instance in the widely used Viterbi algorithm for hidden Markov models [24], in max-product belief propagation [23], in optimal control methods [3] and in resource allocation problems [2].”
Dynamic Computation Offloading for Mobile-Edge Computing With Energy Harvesting Devices (2016, background): “, the relative value iteration algorithm and the linear programming reformulation approach [39].”
Superhuman performance of surgical tasks by robots using iterative learning from human-guided demonstrations (2010, methodology): “We use an LQR controller [3], [6], [16] to execute the trajectory, for it gives the optimal control policy given a linear dynamics model and a quadratic cost function (which penalizes deviations from the trajectory and non-smoothnesses in its execution).”
Q-learning and enhanced policy iteration in discounted dynamic programming (2010, methodology): “Regarding (a) above, note that aside from their conceptual/analytical value, lookup table representation methods can be applied to large scale problems through the use of aggregation (a low-dimensional aggregate representation of a large, possibly inﬁnite-dimensional problem; see Jaakkola, Jordan, a”
Distributed off-Policy Actor-Critic Reinforcement Learning with Policy Consensus (2019, methodology): “Tabular-based methods are generally easier to analyze [3], however, they require the state and action spaces to be discrete and finite.”
Energy Harvesting Wireless Communications: A Review of Recent Advances (2015, methodology): “The long-term average reward per slot can be computed as and the optimal strategy can be numerically found using standard techniques, such as the Policy Iteration Algorithm (PIA) [78].”
Distributed Policy Evaluation Under Multiple Behavior Strategies (2013, methodology): “The predictions by the agents are made in the form of value functions [2], [6], [7].”