Parallelization of the Volterra algorithm for linear optimal open loop control
If and when the time variation of optimal controls of a linear system subject to known forces is required, they can be obtained by the computationally advantageous open loop Volterra formulation (as opposed to the costlier Riccati formulation). For the computation, the Volterra equation is discretized in the time domain via such schemes as trapezoidal integration or SIMPSON's rule and the resulting linear system is solved to obtain the control vector values at discrete time points within the control time T. In the case of very large order systems (degrees of freedom ∼ 5000) a parallel technique is absolutely neccessary, and this paper enunciates an efficient parallel stratagem with efficiencies in the range of 80% and 100%. The algorithm uses 's + 1' processors, 's' being the number of intervals within the control time T, and typically each processor characterizes one time point. © 1989 Springer-Verlag.
Das, SK; Utku, S; Salama, M
Volume / Issue
Start / End Page
Electronic International Standard Serial Number (EISSN)
International Standard Serial Number (ISSN)
Digital Object Identifier (DOI)