• Corpus ID: 54645973

A Viscosity Approach to Stochastic Differential Games of Control and Stopping Involving Impulsive Control

@article{Mguni2018AVA,
  title={A Viscosity Approach to Stochastic Differential Games of Control and Stopping Involving Impulsive Control},
  author={David Henry Mguni},
  journal={arXiv: Optimization and Control},
  year={2018}
}
  • D. Mguni
  • Published 30 March 2018
  • Mathematics
  • arXiv: Optimization and Control
This paper analyses a stochastic differential game of control and stopping in which one of the players modifies a diffusion process using impulse controls, an adversary then chooses a stopping time to end the game. The paper firstly establishes the regularity and boundedness of the upper and lower value functions from which an appropriate variant of the dynamic programming principle (DPP) is derived. It is then proven that the upper and lower value functions coincide so that the game admits a… 
Fault-Tolerant Reinforcement Learning in Continuous Time
  • D. Mguni
  • Computer Science, Mathematics
  • 2019
TLDR
This work proposes a new framework that enables an RL agent to learn controls that are robust against faults and random stoppages in worst case scenarios suitable for applications with continuous state and action spaces.
Cutting Your Losses: Learning Fault-Tolerant Control and Optimal Stopping under Adverse Risk
TLDR
This work proposes a novel approach to risk minimisation within RL in which, in addition to taking actions that maximise its expected return, the controller learns a policy that is robust against stoppages due to an adverse event such as an abrupt failure.
Learning to Shape Rewards using a Game of Switching Controls
TLDR
It is proved that ROSA, which easily adopts existing RL algorithms, learns to construct a shapingreward function that is tailored to the task thus ensuring efficient convergence to high performance policies.
Learning to Shape Rewards using a Game of Two Partners
TLDR
It is proved that ROSA, which adopts existing RL algorithms, learns to construct a shaping-reward function that beneficial to the task thus ensuring e-cient convergence to high performance policies.
DESTA: A Framework for Safe Reinforcement Learning with Markov Games of Intervention
TLDR
A new generation of RL solvers that learn to minimise safety violations while maximising the task reward to the extent that can be tolerated by the safe policy.
SEREN: Knowing When to Explore and When to Exploit
TLDR
It is proved that SEREN converges quickly and induces a natural schedule towards pure exploitation, and can be readily combined with existing RL algorithms to yield improvement in performance relative to state-of-the-art algorithms.
Timing is Everything: Learning to Act Selectively with Costly Actions and Budgetary Constraints
TLDR
This paper proves that LICRA, which seamlessly adopts any RL method, converges to policies that optimally select when to perform actions and their optimal magnitudes and shows LICRA learns the optimal value function and ensures budget constraints are satisfied almost surely.
LIGS: Learnable Intrinsic-Reward Generation Selection for Multi-Agent Learning
TLDR
A new general framework for improving coordination and performance of multi-agent reinforcement learners (MARL), named Learnable Intrinsic-Reward Generation Selection algorithm (LIGS), which introduces an adaptive learner, Generator that observes the agents and learns to construct intrinsic rewards online that coordinate the agents’ joint exploration and joint behaviour.

References

SHOWING 1-10 OF 38 REFERENCES
Stochastic Differential Games Involving Impulse Controls and Double-Obstacle Quasi-variational Inequalities
  • A. Cosso
  • Mathematics
    SIAM J. Control. Optim.
  • 2013
TLDR
It is proved that the upper and lower value functions coincide, indeed it is shown, by means of the dynamic programming principle for the stochastic differential game, that they are the unique viscosity solution to the HJBI equation, therefore proving that the game admits a value.
On the Multi-Dimensional Controller-and-Stopper Games
TLDR
Under appropriate conditions, it is shown that the game has a value and the value function is the unique viscosity solution to an obstacle problem for a Hamilton-Jacobi-Bellman equation.
Impulse Control of Multidimensional Jump Diffusions in Finite Time Horizon
TLDR
This paper establishes rigorously an appropriate form of the dynamic programming principle and shows that the value function is a viscosity solution for the associated Hamilton--Jacobi--Bellman equation involving integro-differential operators.
Zero-sum differential games involving impulse controls
In this paper we are concerned with zero-sum differential games with impulse controls, as well as continuous and switching controls. The motivation is optimal impulse control problems with
Optimal Stochastic Impulse Control with Delayed Reaction
Abstract We study impulse control problems of jump diffusions with delayed reaction. This means that there is a delay δ>0 between the time when a decision for intervention is taken and the time when
On the Existence of Solutions to a Differential Game
In this paper we consider the problem of the existence of a “min-sup” strategy to a pursuit-evasion game. The dynamics of the players have been modeled by a general dynamical system rather than by a
A double obstacle problem arising in differential game theory
Impulse Control of Multidimensional Jump Diffusions
TLDR
Surprisingly, despite these jumps, the regularity properties of the value function for an infinite-horizon discounted cost impulse control problem obtain the same degree of regularity as for the diffusion case, at least when the jump satisfies certain integrability conditions.
Finite Horizon Optimal Stopping of Time-Discontinuous Functionals with Applications to Impulse Control with Delay
TLDR
This work constructs $\varepsilon$-optimal stopping times and provides conditions under which an optimal stopping time exists and demonstrates how to approximate this optimal stopped time by solutions to discrete-time problems.
On Dynkin games with incomplete information
TLDR
It is shown that these games have a value which can be characterized as a viscosity solution to a fully non-linear variational PDE and derive a dual representation of the value function in terms of a minimization procedure.
...
...