Dual heuristic programming for fuzzy control

Lendaris, George G.; Shannon, T.T.; Schultz, L.; Hutsell, S.; Rogers, Anne

doi:10.1109/nafips.2001.944312

Cited by 10 publications

(10 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Dual Heuristic Programming (DHP), critic's output is a derivative of the value function with respect to the state. DHP has been shown to converge to the optimal solution more rapidly than HDP [13], but use of derivatives leads to a complex relationship for updating control and value functional derivatives and poses a function approximation challenge. To overcome these difficulties, Global Dual Heuristic Programming (GDHP) has been proposed that combines advantages of both DHP and HDP [14].…”

Section: Model Based Rlmentioning

confidence: 99%

“…Design difficulties of Markov game based control 1. A nontrivial design consideration is the use of LP in the proposed approach for solving the control-disturber game to update the Q-values (13). In general, LP converges slowly which implies that computational complexity of the approach would increase with the dimensionality of the state-action space.…”

Section: Game Theory Based Solution To Dec-pomdpmentioning

confidence: 99%

See 1 more Smart Citation

Synergizing reinforcement learning and game theory—A new direction for control

Sharma

Gopal

2010

Applied Soft Computing

View full text Add to dashboard Cite

Section: Model Based Rlmentioning

confidence: 99%

Section: Game Theory Based Solution To Dec-pomdpmentioning

confidence: 99%

Synergizing reinforcement learning and game theory—A new direction for control

Sharma

Gopal

2010

Applied Soft Computing

View full text Add to dashboard Cite

“…A family of ADP structures was proposed by Werbos in the early 1990's [68], [69], and has been widely used by others [16], [17], [18], [19], [20], [23], [24], [25], [26], [27], [28], [36], [45], [46], [47], [51], [53], [54]. While the original formulation was based on neural network implementations, it was noted that any learning structure capable of implementing the appropriate mathematics would work.…”

Section: Adaptive Critics To Adpmentioning

confidence: 99%

“…While the original formulation was based on neural network implementations, it was noted that any learning structure capable of implementing the appropriate mathematics would work. Fuzzy Logic structures would be a case in point; examples may be found in [27], [52], [57], [61], [63]. Werbos' family (also called "ladder") of ADP structures includes: Heuristic Dynamic Programming (HDP), Dual Heuristic Programming (DHP), and Global Dual Heuristic Programming (GDHP).…”

Section: Adaptive Critics To Adpmentioning

confidence: 99%

A retrospective on Adaptive Dynamic Programming for control

Lendaris

2009

2009 International Joint Conference on Neural Networks

View full text Add to dashboard Cite

Some three decades ago, certain computational intelligence methods of reinforcement learning were recognized as implementing an approximation of Bellman's Dynamic Programming method, which is known in the controls community as an important tool for designing optimal control policies for nonlinear plants and sequential decision making. Significant theoretical and practical developments have occurred within this arena, mostly in the past decade, with the methodology now usually referred to as Adaptive Dynamic Programming (ADP). The objective of this paper is to provide a retrospective of selected threads of such developments. In addition, a commentary is offered concerning present status of ADP, and threads for future research and development within the controls field are suggested.

show abstract

“…Simple gradient descent was used to train the critic and controller consequent parameters only (see [6] [11] for details on the DHP algorithm and training equations for fuzzy structures). Perturbation through the plant model was used to obtain plant derivatives.…”

Section: Dhp Setup and Training Processmentioning

confidence: 99%