# Per-run Algorithm Selection with Warm-starting using Trajectory-based Features

arXiv: https://arxiv.org/abs/2204.09483 · HTML: https://mlsanity.com/paper/2204.09483/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Apr 2022.

## Built on by (most cited first, 20 of 20)

- [Toward Automated Algorithm Design: A Survey and Practical Guide to Meta-Black-Box-Optimization](https://mlsanity.com/paper/2411.00625/) (2024)
- [To Switch or not to Switch: Predicting the Benefit of Switching between Algorithms based on Trajectory Features](https://mlsanity.com/paper/2302.09075/) (2023)
- [DynamoRep: Trajectory-Based Population Dynamics for Classification of Black-box Optimization Problems](https://mlsanity.com/paper/2306.05438/) (2023)
- [On the Utility of Probing Trajectories for Algorithm-Selection](https://mlsanity.com/paper/2401.12745/) (2024)
- [MA-BBOB: Many-Affine Combinations of BBOB Functions for Evaluating AutoML Approaches in Noiseless Numerical Black-Box Optimization Contexts](https://mlsanity.com/paper/2306.10627/) (2023)
- [Using Automated Algorithm Configuration for Parameter Control](https://mlsanity.com/paper/2302.12334/) (2023)
- [PS-AAS: Portfolio Selection for Automated Algorithm Selection in Black-Box Optimization](https://mlsanity.com/paper/2310.10685/) (2023)
- [Assessing the Generalizability of a Performance Predictive Model](https://mlsanity.com/paper/2306.00040/) (2023)
- [Generalization Ability of Feature-based Performance Prediction Models: A Statistical Analysis across Benchmarks](https://mlsanity.com/paper/2405.12259/) (2024)
- [RF+clust for Leave-One-Problem-Out Performance Prediction](https://mlsanity.com/paper/2301.09524/) (2023)
- [A Survey of Features Used for Representing Black-box Single-objective Continuous Optimization](https://mlsanity.com/paper/2406.06629/) (2024)
- [Towards Automated Design of Bayesian Optimization via Exploratory Landscape Analysis](https://mlsanity.com/paper/2211.09678/) (2022)
- [Sensitivity Analysis of RF+clust for Leave-one-problem-out Performance Prediction](https://mlsanity.com/paper/2305.19375/) (2023)
- [Improving Algorithm-Selection and Performance-Prediction via Learning Discriminating Training Samples](https://mlsanity.com/paper/2404.05359/) (2024)
- [Identifying Easy Instances to Improve Efficiency of ML Pipelines for Algorithm-Selection](https://mlsanity.com/paper/2406.16999/) (2024)
- [Explainable Model-specific Algorithm Selection for Multi-Label Classification](https://mlsanity.com/paper/2211.11227/) (2022)
- [Algorithm Selection with Probing Trajectories: Benchmarking the Choice of Classifier Model](https://mlsanity.com/paper/2501.11414/) (2025)
- [Does Dimensionality Reduction via Random Projections Preserve Landscape Features?](https://mlsanity.com/paper/2604.13230/) (2026)
- [On the Influence of the Feature Computation Budget on Per-Instance Algorithm Selection for Black-Box Optimization](https://mlsanity.com/paper/2605.04954/) (2026)
- [Reinforcement learning to choose optimizers](https://mlsanity.com/paper/2609.01811/) (2026)

## Abstract (arXiv)

Per-instance algorithm selection seeks to recommend, for a given problem instance and a given performance criterion, one or several suitable algorithms that are expected to perform well for the particular setting. The selection is classically done offline, using openly available information about the problem instance or features that are extracted from the instance during a dedicated feature extraction step. This ignores valuable information that the algorithms accumulate during the optimization process. In this work, we propose an alternative, online algorithm selection scheme which we coin per-run algorithm selection. In our approach, we start the optimization with a default algorithm, and, after a certain number of iterations, extract instance features from the observed trajectory of this initial optimizer to determine whether to switch to another optimizer. We test this approach using the CMA-ES as the default solver, and a portfolio of six different optimizers as potential algorithms to switch to. In contrast to other recent work on online per-run algorithm selection, we warm-start the second optimizer using information accumulated during the first optimization phase. We show that our approach outperforms static per-instance algorithm selection. We also compare two different feature extraction principles, based on exploratory landscape analysis and time series analysis of the internal state variables of the CMA-ES, respectively. We show that a combination of both feature sets provides the most accurate recommendations for our test cases, taken from the BBOB function suite from the COCO platform and the YABBOB suite from the Nevergrad platform.