Add citations for NLopt and pagmo2 libraries

Author: GermanHeim

paper/paper.bib

@@ -133,4 +133,27 @@ @misc{cobyla
   url       = {https://github.com/relf/cobyla/},
   version   = {0.7.0},
   year      = {2021}
+}
+
+@article{Biscani2020,
+  author    = {Francesco Biscani and Dario Izzo},
+  doi       = {10.21105/joss.02338},
+  journal   = {Journal of Open Source Software},
+  number    = {53},
+  pages     = {2338},
+  publisher = {The Open Journal},
+  title     = {A parallel global multiobjective framework for optimization: pagmo},
+  url       = {https://doi.org/10.21105/joss.02338},
+  volume    = {5},
+  year      = {2020}
+}
+
+@misc{nlopt,
+  author    = {Steven G. Johnson},
+  journal   = {GitHub repository},
+  publisher = {GitHub},
+  title     = {The {NLopt} nonlinear-optimization package},
+  url       = {http://github.com/stevengj/nlopt},
+  version   = {2.10.0},
+  year      = {2007}
 }

paper/paper.md

@@ -34,6 +34,8 @@ Global optimization is a fundamental problem in engineering, physics, and applie
 
 Building on these algorithmic advances, practical software implementations have become essential for researchers and engineers. MATLAB’s `GlobalSearch`, for example, has been widely adopted, with applications ranging from airfoil design optimization [@airfoil] to kinetic mechanism analysis of polyurethane decomposition [@kinetics]. However, despite its popularity, to the best of our knowledge, no actively maintained open-source alternative currently exists. The only open-source implementation, `MEIGO` [@meigo], is still written in MATLAB, a proprietary language, limiting accessibility.
 
+Several high-quality open-source optimization libraries are available, such as `NLopt` [@nlopt] and `pagmo2` [@Biscani2020], both of which provide Rust and Python bindings. These libraries offer a broad collection of local and global optimization algorithms. However, neither of them provides an implementation of the `OQNLP` algorithm specifically.
+
 `GlobalSearch-rs` addresses this gap and extends its functionality with additional features designed to improve practical usability. It offers a modern, high-performance implementation for bound-constrained nonlinear programming (NLP) problems:
 
 $$