Nonlinear Programming Codes

Information, Tests, Performance

I: Introduction.
II: Optimization methods.
1. Line-search algorithms.
2. Quadratic programming.
3. Unconstrained optimization.
4. Penalty methods.
5. Multiplier methods.
6. Quadratic approximation methods.
7. Generalized reduced gradient methods.
8. The method of Robinson.
III: Optimization programs.
1. Program organization.
2. Description of the programs.
IV: The construction of test problems.
1. Fundamentals of the test problem generator.
2. General test problems.
3. Linearly constrained test problems.
4. Degenerate test problems.
5. Ill-conditioned test problems.
6. Indefinite test problems.
7. Convex test problems.
V: Performance evaluation.
1. Notations.
2. Efficiency, reliability, and global convergence.
3. Performance for solving degenerate, ill-conditioned, and indefinite problems.
4. Sensitivity to slight variations of the problem.
5. Sensitivity to the position of the starting point.
6. Ease of use.
7. How to get a final score.
VI: Conclusions, recommendations, remarks.
1. Pinal conclusions.
2. Recommendations for the design of optimization programs.
3. Some technical details.
Appendix A : Numerical data for constructing test problems.
Appendix B : Sensitivity analysis for the test problems.
Appendix C : Further results.
Appendix D : Evaluation of significance factors.
References.