LumiX: A type-safe, data-centric python library for modern mathematical optimization

Siyah B., Berber T.

SoftwareX, cilt.33, 2026 (SCI-Expanded, Scopus) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 33
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1016/j.softx.2026.102533
  • Dergi Adı: SoftwareX
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, INSPEC, Directory of Open Access Journals
  • Anahtar Kelimeler: Mathematical optimization, Operational research, Python, Sensitivity analysis
  • Karadeniz Teknik Üniversitesi Adresli: Evet

Özet

This paper presents LumiX, an open-source Python library for mathematical optimization designed for data-intensive applications. LumiX employs a data-centric, type-safe modeling paradigm in which problem data, scenario parameters, and optimization models are managed within a unified framework. Key features include Object-Relational Mapping (ORM) integration for automatic variable generation, a solver-agnostic API supporting OR-Tools, Gurobi, CPLEX, and GLPK, automatic linearization of common non-linear expressions, native goal programming, and integrated analysis tools for sensitivity, scenario, and what-if analyses. We present a multi-stage timetabling case study and a quantitative benchmark comparing LumiX against Pyomo and PuLP. The evaluation demonstrates LumiX’s position as a framework for researchers and practitioners developing data-driven optimization solutions, addressing the gap between lightweight procedural libraries and traditional Algebraic Modeling Languages (AMLs). Current limitations, including Big-M parameter sensitivity and McCormick relaxation tightness, are discussed.