Differential Dynamic Programming |
From inside the book
Results 1-3 of 46
Page 8
... yields the following difference equation for the optimal cost of a discrete - time system : Vk ° ( xx ) = min [ Lx ( xx , Ux ) + Vx + 1 ( xx + 1 ) ] uk ( 1.4.17 ) We assume that a nominal control sequence { u : k = 1 , ... , N - 1 } and ...
... yields the following difference equation for the optimal cost of a discrete - time system : Vk ° ( xx ) = min [ Lx ( xx , Ux ) + Vx + 1 ( xx + 1 ) ] uk ( 1.4.17 ) We assume that a nominal control sequence { u : k = 1 , ... , N - 1 } and ...
Page 185
... yields zero variance estimates of the control parameters required in Sections 6.2.1 to 6.2.3 . Note that , for an LQS system , application of a naive Monte - Carlo technique yields zero variance estimates of the second - order ...
... yields zero variance estimates of the control parameters required in Sections 6.2.1 to 6.2.3 . Note that , for an LQS system , application of a naive Monte - Carlo technique yields zero variance estimates of the second - order ...
Page 187
... yield a zero variance estimate of EÃ ( t ) . ] 6.5.4 . Control Variate Technique To employ the control variate technique of variance reduction , an model of the original process has to be obtained . Assume therefore that we have ...
... yield a zero variance estimate of EÃ ( t ) . ] 6.5.4 . Control Variate Technique To employ the control variate technique of variance reduction , an model of the original process has to be obtained . Assume therefore that we have ...
Contents
with Control Inequality Constraints | 11 |
New Algorithms for the Solution of a Class | 21 |
References | 70 |
Copyright | |
5 other sections not shown
Common terms and phrases
assume B₁ bang-bang control Bellman boundary conditions calculated Calculus of Variations Chapter continuous-time system control function control law cost function D. H. Jacobson D. Q. Mayne denote described difference equations differential dynamic programming differential equations discrete-time dx(t Equa error estimate f(x+dx First-Order Algorithm given by Equation initial condition integration interval Lagrange multiplier linear LQP problem McReynolds and Bryson minimizing Mitter N₁ nominal trajectory nonlinear nonoptimal obtained open-loop control optimal control optimal control problems optimal cost optimal trajectory parameters positive-definite random variables reduction in cost Riccati equation S. E. Dreyfus satisfied second-order algorithm sequence side of Equation solution stochastic Sufficient conditions sufficiently small switch t₁ t₂ teff tion ū+du u₁ Univ unspecified arguments V₁ V₂ Vxx fx Vxxdx x₁ Xi+1 yields zero δι δχ นน