ConceptAI & ML
Optimal control
The more general problem of selecting controls that optimize some reward or minimize some cost.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Cost function1Engineering1Markov decision processes1Reward function1System dynamics1Transition dynamics1
Opens up
Concepts that name this one as a prerequisite.
Bang-bang control2Dynamic programming2Hamilton-Jacobi-Bellman (HJB) equation2Inverse reinforcement learning2Trajectory optimization2Bellman value function1Bellman's optimality principle1Certainty equivalent model predictive control1Constrained optimal control problem1Cost-to-go as distance metric1Cost-to-go function1Differential flatness1
Lecturers
2Each in their own terms, from the easiest level.
David Silver
Studying the same type of problems with many of the same methods under different names, which is how to optimally decide a sequence of actions so as to get the best results at the end of the day.Sergey Levine
Lectures that assume it
17The first 12 are shown.
- Commodity Models
- Dynamic Programming I
- Dynamic Programming II
- Computing Lyapunov Functions I
- Convex Optimization II: Model Predictive Control
- Convex Optimization II: Stochastic Model Predictive Control and Branch-and-Bound Methods
- Inverse Reinforcement Learning, Part 1
- Objective-Driven AI: World Models that Learn, Remember, Reason, and Plan
- Koopman Operator Theory and Linear Representations of Nonlinear Dynamics
- Model-Based Reinforcement Learning with Policies
- Trajectory Optimization I
- Trajectory Stabilization
Connections
28The lectures that define it or assume it, and the concepts next to it.
Concepts8
- Cost function
- Markov decision processes
- Reward function
- Bang-bang control
- Dynamic programming
- Hamilton-Jacobi-Bellman (HJB) equation
Show 2 moreShow fewer
Lectures20
- Introduction to Reinforcement Learning
- Foundational concepts in deep reinforcement learning
- Control as Inference, Part 1
- Commodity Models
- Dynamic Programming I
- Dynamic Programming II
Show 14 moreShow fewer
- Computing Lyapunov Functions I
- Convex Optimization II: Model Predictive Control
- Convex Optimization II: Stochastic Model Predictive Control and Branch-and-Bound Methods
- Inverse Reinforcement Learning, Part 1
- Objective-Driven AI: World Models that Learn, Remember, Reason, and Plan
- Koopman Operator Theory and Linear Representations of Nonlinear Dynamics
- Model-Based Reinforcement Learning with Policies
- Trajectory Optimization I
- Trajectory Stabilization
- Mixed-discrete (combinatorial) and continuous optimization
- Stochastic Control
- Deep Inference with Karl J. Friston
- Sampling-based (kinodynamic) motion planning
- Bridging Rigorous Methods with Machine Learning in Robotics