论文信息 - A Parallel Projected Fletcher-Reeves Method for Optimal Control

A Parallel Projected Fletcher-Reeves Method for Optimal Control

We propose a new parametrized gradient projection algorithm for solving large scale constrained optimization problems, and we demonstrate thlie efficiency of the approach by comparing implementations on both the Cray 2 and Connection Machine 2. Numerical results are provided in solving discrete optimal control problems with very high dimensionality (up to 2,097,152 variables).

Gerard G. L. Meyer | Louis J. Podrazik

[1] C. M. Reeves,et al. Function minimization by conjugate gradients , 1964, Comput. J..

[2] D. Bertsekas. Projected Newton methods for optimization problems with simple constraints , 1981, 1981 20th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.

[3] D. Bertsekas. On the Goldstein-Levitin-Polyak gradient projection method , 1974, CDC 1974.

[4] Philip Wolfe,et al. An algorithm for quadratic programming , 1956 .

[5] Stephen J. Wright. Solution of discrete-time optimal control problems on parallel computers , 1990, Parallel Comput..

[6] A. Goldstein. Convex programming in Hilbert space , 1964 .

[7] J. B. Rosen. The Gradient Projection Method for Nonlinear Programming. Part I. Linear Constraints , 1960 .

[8] Stephen J. Wright. Convergence of SQP-like methods for constrained optimization , 1989 .

[9] Louis J. Podrazik,et al. Parallel Recurrence Solvers for Vector and SIMD Supercomputers , 1992, J. Parallel Distributed Comput..

[10] Louis J. Podrazik. Parallel implementations of gradient based iterative algorithms for optimal control problems , 1988 .