VOPy: A Framework for Black-box Vector Optimization
Yaşar Cahit Yıldırım, Efe Mert Karagözlü, İlter Onat Korkmaz, Çağın Ararat, Cem Tekin
TL;DR
This paper presents VOPy, an open-source Python framework for black-box vector optimization where a cone-induced partial order $\preceq_C$ on $\mathbb{R}^D$ defines dominance among vector-valued objectives. It introduces a modular architecture with four core interfaces (Order, Model, Algorithm, Problem) and a ConfidenceRegion abstraction (Rectangular and Ellipsoidal) that enable robust, uncertainty-aware comparisons via CVXPY. The library ships with ready-to-use VO and MOO algorithms (e.g., Naive Elimination, PaVeBa, PaVeBa-GP, VOGP) and supports decoupled evaluations, polyhedral cone definitions, and Gaussian-process-based models implemented with GPyTorch. VOPy is validated through extensive testing, documentation, and example datasets, serving as a flexible platform for researchers to develop, benchmark, and apply cone-based vector optimization methods. As the first open-source toolkit dedicated to black-box VO, VOPy aims to catalyze research and practical applications across noisy, batch, and constrained optimization settings.
Abstract
We introduce VOPy, an open-source Python library designed to address black-box vector optimization, where multiple objectives must be optimized simultaneously with respect to a partial order induced by a convex cone. VOPy extends beyond traditional multi-objective optimization (MOO) tools by enabling flexible, cone-based ordering of solutions; with an application scope that includes environments with observation noise, discrete or continuous design spaces, limited budgets, and batch observations. VOPy provides a modular architecture, facilitating the integration of existing methods and the development of novel algorithms. We detail VOPy's architecture, usage, and potential to advance research and application in the field of vector optimization. The source code for VOPy is available at https://github.com/Bilkent-CYBORG/VOPy.