Robustness of linearly solvable Markov games employing inaccurate dynamics model
نویسندگان
چکیده
منابع مشابه
Linearly-solvable Markov decision problems
Advances in Neural Information Processing Systems 2006 We introduce a class of MPDs which greatly simplify Reinforcement Learning. They have discrete state spaces and continuous control spaces. The controls have the effect of rescaling the transition probabilities of an underlying Markov chain. A control cost penalizing KL divergence between controlled and uncontrolled transition probabilities ...
متن کاملHierarchical Linearly-Solvable Markov Decision Problems
We present a hierarchical reinforcement learning framework that formulates each task in the hierarchy as a special type of Markov decision process for which the Bellman equation is linear and has analytical solution. Problems of this type, called linearly-solvable MDPs (LMDPs) have interesting properties that can be exploited in a hierarchical setting, such as efficient learning of the optimal ...
متن کاملHierarchy through Composition with Linearly Solvable Markov Decision Processes
Hierarchical architectures are critical to the scalability of reinforcement learning methods. Current hierarchical frameworks execute actions serially, with macroactions comprising sequences of primitive actions. We propose a novel alternative to these control hierarchies based on concurrent execution of many actions in parallel. Our scheme uses the concurrent compositionality provided by the l...
متن کاملLinearly Solvable Optimal Control
We summarize the recently-developed framework of linearly-solvable stochastic optimal control. Using an exponential transformation, the (Hamilton-Jacobi) Bellman equation for such problems can bemade linear, giving rise to efficient numericalmethods. Extensions to game theory are also possible and lead to linear Isaacs equations. The key restriction that makes a stochastic optimal control probl...
متن کاملLinearly Equivalent Actions of Solvable Groups
Let G be a group. A G-set is a finite set provided with a left action of g 4 G. For a G-set X and g g G we write X s x g X : gx s x . The permutation character of a G-set X is the function p from G to the ring X Z of integers that sends g to the cardinality aX g of X . We say that two G-sets X and Y are linearly equï alent if p s p . By representation X Y theory of finite groups, two G-sets X a...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Artificial Life and Robotics
سال: 2017
ISSN: 1433-5298,1614-7456
DOI: 10.1007/s10015-017-0401-2