From Performance to Understanding: A Vision for Explainable Automated Algorithm Design

TL;DR

This paper advocates for integrating explainability into automated algorithm design using systematic benchmarking, LLM-driven discovery, and problem descriptors to foster interpretable, class-specific optimization strategies.

Contribution

It proposes a comprehensive framework combining automation and understanding through benchmarking, discovery, and problem analysis to advance explainable algorithm design.

Findings

Framework for explainable algorithm design outlined

Benchmarking attributes performance to components

Problem descriptors connect behavior to landscape structure

Abstract

Automated algorithm design is entering a new phase: Large Language Models can now generate full optimisation (meta)heuristics, explore vast design spaces and adapt through iterative feedback. Yet this rapid progress is largely performance-driven and opaque. Current LLM-based approaches rarely reveal why a generated algorithm works, which components matter or how design choices relate to underlying problem structures. This paper argues that the next breakthrough will come not from more automation, but from coupling automation with understanding from systematic benchmarking. We outline a vision for explainable automated algorithm design, built on three pillars: (i) LLM-driven discovery of algorithmic variants, (ii) explainable benchmarking that attributes performance to components and hyperparameters and (iii) problem-class descriptors that connect algorithm behaviour to landscape structure. Together, these elements form a closed knowledge loop in which discovery, explanation and generalisation reinforce each other. We argue that this integration will shift the field from blind search to interpretable, class-specific algorithm design, accelerating progress while producing reusable scientific insight into when and why optimisation strategies succeed.