Towards Dynamic Algorithm Selection for Numerical Black-Box Optimization: Investigating BBOB as a Use Case
This addresses the algorithm selection problem in evolutionary computation, which is incremental as it builds on existing methods for dynamic configuration.
The paper tackles the problem of selecting the best algorithm for black-box optimization by proposing dynamic algorithm selection (dynAS), which switches between algorithms during a run, and shows that even single-switch dynAS can lead to significant performance gains using the BBOB benchmark.
One of the most challenging problems in evolutionary computation is to select from its family of diverse solvers one that performs well on a given problem. This algorithm selection problem is complicated by the fact that different phases of the optimization process require different search behavior. While this can partly be controlled by the algorithm itself, there exist large differences between algorithm performance. It can therefore be beneficial to swap the configuration or even the entire algorithm during the run. Long deemed impractical, recent advances in Machine Learning and in exploratory landscape analysis give hope that this dynamic algorithm configuration~(dynAC) can eventually be solved by automatically trained configuration schedules. With this work we aim at promoting research on dynAC, by introducing a simpler variant that focuses only on switching between different algorithms, not configurations. Using the rich data from the Black Box Optimization Benchmark~(BBOB) platform, we show that even single-switch dynamic Algorithm selection (dynAS) can potentially result in significant performance gains. We also discuss key challenges in dynAS, and argue that the BBOB-framework can become a useful tool in overcoming these.