NEWS.md

In SBOAtools: Secretary Bird Optimization for Continuous Optimization and Neural Network Training

SBOAtools 0.1.1

New features

• Added 23 built-in benchmark functions, named F1 to F23, for continuous optimization problems.
• Added list_benchmarks() to display available benchmark functions and their metadata.
• Added get_benchmark() to retrieve benchmark definitions, bounds, dimensions, categories, and labels.

Improvements

• Updated sboa() so that built-in benchmark names, such as "F1", can be supplied directly as the objective function.
• Added benchmark metadata, including labels, categories, lower and upper bounds, and fixed dimensions where applicable.
• Improved usability for benchmark-based experiments by allowing users to access standard test functions without manually defining objective functions.

Compatibility

• Preserved compatibility with user-defined objective functions.
• Preserved compatibility with sboa_mlp().
• Existing workflows based on sboa() and sboa_mlp() remain unchanged.

SBOAtools documentation built on May 3, 2026, 9:06 a.m.