Associate Professor (with Habilitation in Mathematics) · Department of Mathematics, ISEL — Polytechnic Institute of Lisbon
Integrated Researcher · IDMEC / LAETA, Instituto Superior Técnico, Universidade de Lisboa
Research in Optimization and Decision Sciences: derivative-free optimization, multiobjective optimization, global and local direct-search methods, and applied operations research developed in partnership with industry.
| Method | Description | Reference |
|---|---|---|
| DMS — Direct MultiSearch | Local multiobjective optimization without derivatives | Custódio, Madeira, Vaz & Vicente, 2011 |
| GLODS — Global and Local Optimization using Direct Search | Single-objective global optimization | Custódio & Madeira, Journal of Global Optimization, 2015 |
| MultiGLODS | Global multiobjective optimization without derivatives | Custódio & Madeira, 2018 |
| GLODS-SI — Scale-Invariant variant for engineering design optimization | Companion code: GLODS_SI |
Madeira, Journal of Computational Design and Engineering (2026), Open Access — DOI 10.1093/jcde/qwag049 |
A two-part contribution on mixed-variable optimisation, in which the structural and algorithmic components reference one another:
- Structural foundation — Mixed-Variable Optimisation as a Metric Product Space: Transient Categorical Geometry and a Hierarchy of Local Optimality. Manuscript under review. Preprint: Zenodo. Companion code:
mixed-spaces-supplementary. - Algorithmic instantiation — Deterministic Neighborhood Rotation (DNR) for Categorical Variables in Derivative-Free Optimization. Working paper. Preprint: Zenodo. Code release planned upon acceptance for publication.
Companion repositories
GLODS_SI— implementation and reproducibility artefacts for GLODS-SIMOO_Prob_Matlab— multiobjective benchmark test problems over continuous and mixed-variable decision spaces under controlled scale heterogeneity · DOI 10.5281/zenodo.20783713DFO_Benchmark_Suite— single-objective derivative-free benchmark over continuous and mixed-variable decision spaces under controlled scale heterogeneity · DOI 10.5281/zenodo.20782892mixed-spaces-supplementary— supplementary code for a manuscript on mixed-variable optimisation as a metric product space
Open, self-contained MATLAB test problems for benchmarking derivative-free solvers under controlled scale heterogeneity, across single- and multi-objective settings and continuous / mixed-variable decision spaces. Each collection instantiates its base problems under 8 deterministic heterogeneity strategies.
| Setting | Variables | Base problems | Instances (× 8 strategies) | Repository |
|---|---|---|---|---|
| Single-objective | Continuous | 63 | 504 | DFO_Benchmark_Suite |
| Single-objective | Mixed (C/D/K) | 63 | 504 | DFO_Benchmark_Suite |
| Multi-objective | Continuous | 108 | 864 | MOO_Prob_Matlab |
| Multi-objective | Mixed (C/D/K) | 108 | 864 | MOO_Prob_Matlab |
| Multi-objective — base set¹ | Continuous | 108 | — (unscaled) | MOO_Prob_Matlab |
The eight strategy families are shared across all collections. The sobol_digit_oscillatory contrast is κ = 1e8 in the single-objective suite and κ = 1e6 in the multi-objective suite, matching the settings used in the respective studies.
¹ Base set: MATLAB conversions of the original Direct MultiSearch (DMS) problems, originally distributed in AMPL — the reference set (problems/) from which the multi-objective collections above are derived.
Archived and citable on Zenodo — DFO_Benchmark_Suite: DOI 10.5281/zenodo.20782892 · MOO_Prob_Matlab: DOI 10.5281/zenodo.20783713.
Derivative-free optimization · Multiobjective optimization · Direct-search methods · Decision Sciences and operations research · Industrial applications in transit operations, vehicle routing, postal logistics and energy systems
ORCID 0000-0001-9523-3808 · CiênciaVitae 6F1E-DCF0-D6EC · ResearcherID N-6918-2016 · Scopus 7003405549 · Personal page
IDMEC · ISEL — Department of Mathematics
h-index 22 · 1,363 citations (Web of Science) · Two PhDs (Mathematics, IST 2019; Mechanical Engineering, IST 2004) · Habilitation in Mathematics (Universidade de Évora, 2021)
- 2023 — IPL / CGD Scientific Excellence Award (Technologies and Engineering, triennium 2020–2022)
- 2021 — IPL / CGD Award for Recognition of Activities with Community Relevance (triennium 2018–2020)
- 2019–2022 — Four IPL / CGD merit diplomas for scientific work in Technologies and Engineering
- 2004–2021 — Excelente (highest rating) in 18 consecutive ISEL teaching evaluations
ISEL · Rua Conselheiro Emídio Navarro 1, 1959-007 Lisboa, Portugal · jose.madeira (at) isel.pt
IDMEC · Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal · aguilarmadeira (at) tecnico.ulisboa.pt
