Algorithms - Fletcher Reeves

The OSTRICH Fletcher Reeves optimization algorithm

Initial Publication:

Modified:

The following optional group will configure the Fletcher-Reeves algorithm and will be processed if [ProgramType] is set to “Fletcher-Reeves”.

BeginFletchReevesAlg
ConvergenceVal [conv_val]
MaxStalls      [max_stalls]
MaxIterations  [max_iter]
EndFletchReevesAlg
BeginFletchReevesAlg
ConvergenceVal 1.00E-6
MaxStalls      3
MaxIterations  20
EndFletchReevesAlg

General Format (left) and Example (right) of the Fletcher-Reeves Group

Where BeginFletchReevesAlg and EndFletchReevesAlg are parsing tags that wrap the following set of algorithm configuration variables:

  • ConvergenceVal, MaxStalls: These variables control the convergence termination criterion for the algorithm. The algorithm will stop when the relative reduction in the objective function over max_stalls iterations is less than conv_val. The default value for conv_val is 1.00E-6 and the default value for max_stalls is 3.
  • MaxIterations:The maximum number of iterations of the algorithm. The default value is 20.

References

Fletcher, R.,Reeves, C. M. 1964. Function minimization by conjugate gradients. The computer journal 7, 149-154.