Abstract:The performance of automated algorithm selection (AAS) strongly depends on the portfolio of algorithms to choose from. Selecting the portfolio is a non-trivial task that requires balancing the trade-off between the higher flexibility of large portfolios with the increased complexity of the AAS task. In practice, probably the most common way to choose the algorithms for the portfolio is a greedy selection of the algorithms that perform well in some reference tasks of interest. We set out in this work to investigate alternative, data-driven portfolio selection techniques. Our proposed method creates algorithm behavior meta-representations, constructs a graph from a set of algorithms based on their meta-representation similarity, and applies a graph algorithm to select a final portfolio of diverse, representative, and non-redundant algorithms. We evaluate two distinct meta-representation techniques (SHAP and performance2vec) for selecting complementary portfolios from a total of 324 different variants of CMA-ES for the task of optimizing the BBOB single-objective problems in dimensionalities 5 and 30 with different cut-off budgets. We test two types of portfolios: one related to overall algorithm behavior and the `personalized' one (related to algorithm behavior per each problem separately). We observe that the approach built on the performance2vec-based representations favors small portfolios with negligible error in the AAS task relative to the virtual best solver from the selected portfolio, whereas the portfolios built from the SHAP-based representations gain from higher flexibility at the cost of decreased performance of the AAS. Across most considered scenarios, personalized portfolios yield comparable or slightly better performance than the classical greedy approach. They outperform the full portfolio in all scenarios.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve a key problem in Automated Algorithm Selection (AAS) in black - box optimization: **How to select an effective algorithm portfolio to balance the flexibility of the portfolio and the complexity of the AAS task**. Specifically, the paper explores the following points: 1. **Challenges in algorithm portfolio selection**: - In black - box optimization, selecting the most suitable algorithm for a specific problem is an important challenge, which depends on the characteristics of the problem instance and the available computing resources. - As the number of available algorithms increases, the AAS task becomes more complex. In practice, there are usually hundreds of different algorithms to choose from. 2. **Limitations of traditional methods**: - The traditional greedy selection method (i.e., selecting algorithms that perform well on some reference tasks) is simple but may lead to a lack of diversity in the portfolio, which in turn affects AAS performance. 3. **Data - driven portfolio selection methods**: - The paper proposes a data - driven method for algorithm portfolio selection. This method creates meta - representations of algorithm behaviors, constructs a similarity graph between algorithms, and applies graph algorithms to select the final portfolio. - Two different meta - representation techniques are studied: performance2vec and SHAP - based representations. 4. **Specific application scenarios**: - The paper uses 324 different variants of the CMA - ES algorithm to optimize BBOB single - objective problems, conducts experiments in 5 - dimensional and 30 - dimensional spaces, and tests the performance under different budgets. 5. **Personalized portfolios**: - The concept of personalized portfolios is proposed, that is, selecting the most appropriate algorithm according to the specific situation of each problem rather than the globally optimal algorithm. 6. **Experimental verification**: - The performance of the proposed portfolio selection method in different scenarios is verified through experiments and compared with the traditional greedy method. In summary, the main goal of the paper is to explore, through a data - driven method, how to select an algorithm portfolio that is both diverse and representative to improve the performance of AAS in black - box optimization.

PS-AAS: Portfolio Selection for Automated Algorithm Selection in Black-Box Optimization

Generalization in portfolio-based algorithm selection

ASlib: A Benchmark Library for Algorithm Selection

A Survey of Meta-features Used for Automated Selection of Algorithms for Black-box Single-objective Continuous Optimization

Extreme Algorithm Selection With Dyadic Feature Representation

Population-based Algorithm Portfolios with Automated Constituent Algorithms Selection

A Portfolio Approach to Algorithm Selection for Discrete Time-Cost Trade-off Problem

Sequential vs. Integrated Algorithm Selection and Configuration: A Case Study for the Modular CMA-ES

Impact of Training Instance Selection on Automated Algorithm Selection Models for Numerical Black-box Optimization

Set-based Particle Swarm Optimization for Portfolio Optimization with Adaptive Coordinate Descent Weight Optimization

AMA-K: Aggressive Multi-Temporal Allocation An Algorithm for Aggressive Online Portfolio Selection

Markowitz-based cardinality constrained portfolio selection using Asexual Reproduction Optimization (ARO)

ParDen: Surrogate Assisted Hyper-Parameter Optimisation for Portfolio Selection

Population-based algorithm portfolios for numerical optimization

Robust and adaptive algorithms for online portfolio selection

Hybridizing Target- and SHAP-encoded Features for Algorithm Selection in Mixed-variable Black-box Optimization

An activity level based surrogate-assisted evolutionary algorithm for many-objective optimization

Comparative Analysis of Selection Hyper-Heuristics for Real-World Multi-Objective Optimization Problems

Using Machine Learning Methods to Assess Module Performance Contribution in Modular Optimization Frameworks

Automatic Algorithm Selection for Pseudo-Boolean Optimization with Given Computational Time Limits

Benchmarking Feature-based Algorithm Selection Systems for Black-box Numerical Optimization