Open AccessProceedings Article
Risk sensitive path integral control
Bart van den Broek,Wim Wiegerinck,Bert Kappen +2 more
- 08 Jul 2010
- pp 615-622
TL;DR: In this article, the authors show that path integral methods generalize directly to risk sensitive stochastic optimal control with non-linear dynamics in continuous space-time, and demonstrate the effect of multi-modal control with risk sensitivity.
read more
Abstract: Recently path integral methods have been developed for stochastic optimal control for a wide class of models with non-linear dynamics in continuous space-time. Path integral methods find the control that minimizes the expected cost-to-go. In this paper we show that under the same assumptions, path integral methods generalize directly to risk sensitive stochastic optimal control. Here the method minimizes in expectation an exponentially weighted cost-to-go. Depending on the exponential weight, risk seeking or risk averse behaviour is obtained. We demonstrate the approach on risk sensitive stochastic optimal control problems beyond the linear-quadratic case, showing the intricate interaction of multi-modal control with risk sensitivity.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Walking and Running on Yielding and Fluidizing Ground
Feifei Qian,Tingnan Zhang,Chen Li,Aaron M. Hoover,Pierangelo Masarati,Paul M. Birkmeyer,Andrew Pullin,Ronald S. Fearing,Daniel I. Goldman +8 more
- 09 Jul 2012
TL;DR: Presented at Robotics: Science and Systems VIII, July 09-July 13, 2012, University of Sydney, Sydney, NSW, Australia.
1K
Active inference: A process theory
Karl J. Friston,Thomas H. B. FitzGerald,Francesco Rigoli,Philipp Schwartenbeck,Giovanni Pezzulo +4 more
TL;DR: The fact that a gradient descent appears to be a valid description of neuronal activity means that variational free energy is a Lyapunov function for neuronal dynamics, which therefore conform to Hamilton’s principle of least action.
Active inference and epistemic value
Karl J. Friston,Francesco Rigoli,Dimitri Ognibene,Christoph Mathys,Thomas H. B. FitzGerald,Giovanni Pezzulo +5 more
TL;DR: A formal treatment of choice behavior based on the premise that agents minimize the expected free energy of future outcomes and ad hoc softmax parameters become the expected (Bayes-optimal) precision of beliefs about, or confidence in, policies.
Active inference and learning
Karl J. Friston,Thomas H. B. FitzGerald,Francesco Rigoli,Philipp Schwartenbeck,John P. O'Doherty,Giovanni Pezzulo +5 more
TL;DR: This work has shown that optimal behaviour is quintessentially belief based, and that habits are learned by observing one’s own goal directed behaviour and selected online during active inference.
593
Active Inference, Curiosity and Insight
TL;DR: This article uses simulations of abstract rule learning and approximate Bayesian inference to show that minimizing (expected) variational free energy leads to active sampling of novel contingencies and closes explanatory gaps in generative models of the world, thereby reducing uncertainty and satisfying curiosity.
References
Space-Time Approach to Non-Relativistic Quantum Mechanics
TL;DR: In this paper, the authors formulated non-relativistic quantum mechanics in a different way and showed that the probability of an event which can happen in several different ways is the absolute square of a sum of complex contributions, one from each alternative way.
•Book
Deterministic and stochastic optimal control
Wendell H. Fleming,Raymond Rishel +1 more
- 17 Nov 1975
TL;DR: In this paper, the authors considered the problem of optimal control of Markov diffusion processes in the context of calculus of variations, and proposed a solution to the problem by using the Euler Equation Extremals.
3.2K
State-space formulae for all stabilizing controllers that satisfy and H ∞ norm bound and relations to risk sensitivity
Keith Glover,John Doyle +1 more
TL;DR: In this paper, the relationship between robust and stochastic control is established, giving an equivalence between robust control and control with respect to the Riccati Equation (RCE).
1.4K
Optimal stochastic linear systems with exponential performance criteria and their relation to deterministic differential games
TL;DR: In this article, two stochastic optimal control problems are solved whose performance criteria are the expected values of exponential functions of quadratic forms, and the optimal controller is linear in both cases but depends upon the covariance matrix of the additive process noise so that the certainty equivalence principle does not hold.
811
Related Papers (5)
Richard S. Sutton,Andrew G. Barto +1 more
- 01 Jan 1988
Laurent Itti,Pierre Baldi +1 more
Matthew J. Beal
- 01 Jan 2003