Artificial intelligence-based automated algorithm discovery: A systematic literature review
Year: 2026
Type: review
Study design: Systematic Review
Abstract: Artificial Intelligence (AI) techniques are increasingly used to design and optimize algorithms, but the research is scattered across several disciplines. This systematic literature review examines AI-based automated algorithm discovery along three dimensions: the problem domains addressed, the AI techniques and search mechanisms that drive the discovery process, and the validation methods used to assess the discovered algorithms. The review follows Kitchenham’s guidelines and reports the selection process in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) format. After selection and quality assessment, 33 primary studies were included. Three main findings emerge. First, the studies concentrate on three problem domains: combinatorial optimization, computational and algebraic foundations, and scientific and symbolic discovery. Second, three technique families dominate: deep reinforcement learning combined with tree search, Large Language Model (LLM)-driven evolutionary search, and hybrid pipelines that combine AI components with classical symbolic engines. Third, the empirical evidence base is limited. Only four of the 33 studies provide mathematical verification, 15 rely only on empirical performance comparisons, two evaluate the discovered algorithms on physical hardware, and one tests whether the discoveries are novel rather than recombinations of patterns learned from the training data. The field therefore needs standardized benchmark suites, protocols that control for memorization, and end-to-end evaluation on physical hardware.
Source: Applied Soft Computing
Institutions: Istanbul Kültür University, Qatar University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,730
Funder: Qatar National Library
OpenAlex ID: W7225788904
