A projected primal-dual gradient optimal control method for deep reinforcement learning
نویسندگان
چکیده
منابع مشابه
Primal-Dual Projected Gradient Algorithms for Extended Linear-Quadratic Programming
Many large-scale problems in dynamic and stochastic optimization can be modeled with extended linear-quadratic programming, which admits penalty terms and treats them through duality. In general the objective functions in such problems are only piecewise smooth and must be minimized or maximized relative to polyhedral sets of high dimensionality. This paper proposes a new class of numerical met...
متن کاملAccelerated Primal-Dual Policy Optimization for Safe Reinforcement Learning
Constrained Markov Decision Process (CMDP) is a natural framework for reinforcement learning tasks with safety constraints, where agents learn a policy that maximizes the long-term reward while satisfying the constraints on the long-term cost. A canonical approach for solving CMDPs is the primal-dual method which updates parameters in primal and dual spaces in turn. Existing methods for CMDPs o...
متن کامل5 A Primal-Dual Active-Set Multigrid Method for Control-Constrained Optimal Control Problems
In this chapter we consider optimal control problems with additional inequality constraints imposed on the control unknown u and for their efficient solution we combine a primal-dual active-set strategy with the multigrid method developed in the previous chapter. Control-constraints are specified by the condition u ∈ Uad, where the set of admissible controls Uad ⊂ L(Ω) is a proper subset of L(Ω...
متن کاملDual fast projected gradient method for quadratic programming
The application of the fast gradient method to the dual QP leads to the Dual Fast Projected Gradient (DFPG) method. The DFPG converges with O ( k−2 ) rate, where k > 0 is the number of steps. At each step, it requires O(nm) operations. Therefore for a given ε > 0 an ε-approximation to the optimal dual function value
متن کاملThe Primal-Dual Hybrid Gradient Method for Semiconvex Splittings
This paper deals with the analysis of a recent reformulation of the primal-dual hybrid gradient method, which allows one to apply it to nonconvex regularizers. Particularly, it investigates variational problems for which the energy to be minimized can be written as G(u) + F (Ku), where G is convex, F is semiconvex, and K is a linear operator. We study the method and prove convergence in the cas...
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: Journal of Mathematics in Industry
سال: 2020
ISSN: 2190-5983
DOI: 10.1186/s13362-020-00075-3