Optimization benchmarks

Mittelmann benchmark

Uno v2.9.0 solves 45/47 instances in the Mittelmann benchmark, in about the same time as IPOPT.

CUTE benchmark

Uno presets have been tested against state-of-the-art solvers on 429 small problems of the CUTE benchmark translated to AMPL. The figure below (dated September 9, 2026) is a performance profile of Uno (solid lines) vs state-of-the-art solvers filterSQP and IPOPT (dotted lines); it shows how many problems are solved for a given budget of function evaluations (1 time, 2 times, 4 times, ..., \(2^x\) times the number of objective evaluations of the best solver for each instance).

The performance profile below features the state-of-the-art solvers SNOPT, MINOS, LANCELOT, LOQO and CONOPT.

All log files can be found here.