Benchmarking Derivative-Free Optimization Algorithms

Stefan Wild (smw58cornell.edu)
Jorge Moré (moremcs.anl.gov)

Abstract: We propose data profiles as a tool for analyzing the performance of derivative-free optimization solvers when there are constraints on the computational budget. We use performance and data profiles, together with a convergence test that measures the decrease in function value, to analyze the performance of three solvers on sets of smooth, noisy, and piecewise-smooth problems. Our results provide estimates for the performance difference between these solvers, and show that on these problems, a model-based solver performs better than geometry-based solvers, even for noisy and piecewise-smooth problems.

Keywords: Derivative-Free Optimization, Benchmarking, Performance Evaluation, Deterministic Simulations, Computational Budget.

Category 1: Nonlinear Optimization (Unconstrained Optimization )

Category 2: Optimization Software and Modeling Systems (Optimization Software Benchmark )

Citation: Preprint ANL/MCS-P1471-1207, Mathematics and Computer Science Division, Argonne National Laboratory, April 2008

Download: [PDF]

Entry Submitted: 01/13/2008
Entry Accepted: 01/13/2008
Entry Last Modified: 05/13/2008

Modify/Update this entry