Benchmarks for Nonlinear Optimization

Benchmarks comparing LOQO with SNOPT (created by Gill, Murray, and Saunders) and NITRO (created by Nocedal, on 800 problems from the CUTE set (assembled by Conn, Gould, and Toint) and from the Schittkowski test set can be found here:


Some tuning was done with respect to each of the codes. For LOQO, the specifics for each tuned problem can be found here:
Click here for the source AMPL models.

See also Hans Mittelman's extensive benchmarks at