This section contains recommendations for optimization algorithms based on problem characteristics.

Unconstrained Optimization Methods

1. No Derivative Information or Noisy Function Evaluations
   • Parallel Direct Search (PDS) Method
2. Derivative Information Available
   • Conjugate Gradient Method
   • Newton's Method

• Barrier Newton Method
• Bound Constrained Ellipsoid Method
• Bound Constrained Newton
• Nonlinear Interior-Point Method
• NPSOL wrapper
• Parallel Direct Search (PDS) Method

• Nonlinear Interior-Point Method
• NPSOL wrapper

Next Section: Frequently Asked Questions | Back to Main Page

Last revised February 4, 2002

Juan Meza meza@ca.sandia.gov



Documentation generated by