This competition challenges participants to design new and interesting benchmark functions for continuous optimization.
Unlike traditional competitions that focus on solving predefined problems, this challenge inverts the paradigm: participants submit problem sets that best differentiate the performance of a predefined set of optimization algorithms.
The objective is to advance our understanding of what makes optimization problems difficult, diverse, or distinctive with respect to algorithm performance.

Participation is simple: you design a set of 25 benchmark problems of the form $0,1^N\rightarrow\mathbb{R}$.
The quality of your problem set is then judged in terms of performance diversity of a set of 5 commonly-used algorithms (CMA-ES, DE, PSO, BFGS, COBYLA). The exact evaluation procedure is present in our getting started files. For details about the limitations on the problem dimensions, details about test instances, etc., please refer to the dedicated competition website.

We particularly welcome submissions that are accompanied by a short paper or technical report (up to 2 pages following the GECCO formatting requirements) describing the key approaches used to create the submitted problems. Each team can submit only a single problem set.