The goal of the Workshop on Automated Heuristic Design and Algorithm Selection (AHDAS) is to provide a specialized forum for researchers and practitioners interested in the automation of algorithmic decision-making for optimization and search. The workshop will focus on hyper-heuristics, automated algorithm selection, algorithm portfolios, automated heuristic generation, automated algorithm configuration, meta-learning, neuroevolution, and learning-based optimization.
The workshop seeks to connect complementary perspectives from artificial intelligence, computational intelligence, operations research, evolutionary computation, machine learning, and automated machine learning. Its purpose is to promote and encourage a critical and nourishing discussion on methods that can automatically design, select, configure, learn, or adapt algorithms according to problem characteristics, instance features, performance feedback, or application requirements.
By bringing this discussion to the Mexican International Conference on Artificial Intelligence (MICAI) 2026, the workshop aims to incentivize collaboration, disseminate recent advances, identify open research challenges, and strengthen the visibility of automated heuristic design and algorithm selection within the broader AI community.
The Workshop on Automated Heuristic Design and Algorithm Selection (ADHAS) invites submissions on methods, models, systems, and applications related to the automation of algorithmic decision-making for optimization, search, and artificial intelligence.
The workshop will be held in conjunction with the 25th Mexican International Conference on Artificial Intelligence, MICAI 2026, which will take place from November 2 to 6, 2026, at Tecnológico de Monterrey, Campus Chihuahua, Mexico. MICAI 2026 is organized by the Mexican Society for Artificial Intelligence and Tecnológico de Monterrey, Campus Chihuahua.
Scope and Motivation
Many real-world computational problems require the use of sophisticated algorithms whose performance depends strongly on the structure of the problem, the characteristics of the instance, and the operational context in which they are deployed. Traditionally, the design, selection, and configuration of these algorithms have relied heavily on expert knowledge, manual experimentation, and problem-specific tuning.
This workshop focuses on approaches that seek to automate these processes. Topics of interest include hyper-heuristics, automated heuristic design, automated algorithm selection, automated algorithm portfolios, automated algorithm configuration, neuroevolution, meta-learning, learning-based optimization, and related areas. The goal is to bring together researchers and practitioners interested in systems that can automatically design, select, configure, or adapt algorithms according to problem features, performance feedback, or changing environments.
The workshop aims to incentivize discussion across artificial intelligence, computational intelligence, operations research, evolutionary computation, machine learning, automated machine learning, and combinatorial optimization.
Topics of Interest
Submissions should address, but are not limited to, the following topics:
- Hyper-heuristics for combinatorial, continuous, dynamic, or multi-objective optimization
- Selection hyper-heuristics and generation hyper-heuristics
- Automated heuristic design and automated algorithm design
- Automated Algorithm selection and per-instance algorithm selection
- Automated Algorithm portfolios and solver portfolios
- Automated Algorithm configuration and parameter control
- Meta-learning for optimization and algorithm recommendation
Instance characterization, feature extraction, and landscape analysis - Learning-based optimization and adaptive search strategies
Reinforcement learning for heuristic selection or algorithm control - Neural, symbolic, evolutionary, or hybrid methods for heuristic generation
- Automated design of metaheuristics and search operators
- Transfer learning and generalization in heuristic or algorithm selection
- Explainability, robustness, and reliability in automated algorithmic systems
- Benchmarking, reproducibility, and performance assessment
- Applications in scheduling, routing, packing, planning, logistics, timetabling, energy systems, bioinformatics, engineering design, and other domains
- Case studies involving real-world algorithm selection, heuristic design, or optimization systems
- Negative results, methodological lessons, and limitations of automated heuristic design
All submissions must be written in English and should present original work that has not been previously published and is not under review elsewhere. MICAI indicates that only complete and finished papers will be reviewed, not abstracts.
Proceedings
Accepted workshop papers will be included in the MICAI proceedings. According to the MICAI 2026 call, workshop proceedings will appear in the Springer LNAI series and are indexed in Scopus, SCImago, DBLP, INSPEC, and EI Compendex.
