I am Yi Zhang, an Optimization Research Scientist at The Optimization Firm working on global optimization and artificial intelligence.
I have been enjoying developing optimization solvers and studying mathematics ever since, and I am still loving every minute of my work. In short, here are the roles that define me:
-
Admirer and developer (2017-Now) of global mixed-integer nonlinear programming (MINLP) solver: BARON
-
Algorithm Expert (2019-2021) in designing optimization and reinforcement learning algorithms for cross-border e-commerce logistics scenarios in Cainiao-Alibaba Group and Lazada Group
-
PhD (2014-2019) in Control Science and Engineering, Zhejiang University, Collaborator (2017-2021) with CAPD of Carnegie Mellon University, with emphasis on modeling, simulation, optimization in process system engineering
-
Researcher that never retires and always enjoys studying
My current research focuses on advancing theories and algorithms in global optimization for Mixed-Integer Nonlinear Programming (MINLP), developing AI-powered optimization software, and enabling data-driven decision-making in manufacturing. Among all the projects I’m working on, now I have a particular interest in The Pooling Problem and Quadratic Unconstrained Binary Optimization (QUBO).
In my research journey (as some might see on Google Scholar), I enjoyed collaborating with outstanding professionals to achieve meaningful breakthroughs, step by step. For more details of my experience, please check my LinkedIn.
News
2025
- Oct New publication:
- Zhang, Y., Sahinidis, N.V. A combined linear and nonlinear presolve for nonlinear optimization. EURO Journal on Computational Optimization (2025).
- Aug Invited talk: Simultaneous convexification for global optimization of nonlinear problems with polynomial functions, GOC 2025.
- Jun New publication:
- Zhang, Y., Sahinidis, N.V. Learning to deactivate probing with graph convolutional network for mixed-integer nonlinear programming. Optim Lett (2025).
- Source code is available for constructing GCN for general learning problems in Mixed-Integer Nonlinear Programming (MINLP)
- Zhang, Y., Sahinidis, N.V. Learning to deactivate probing with graph convolutional network for mixed-integer nonlinear programming. Optim Lett (2025).
- Apr Open-source code:
- Black-box solver performance visualizer (BBOviz). Developed by Zhang, Y. and Sahinidis, N.V.
- A Python script for analyzing and visualizing the performance of black-box solvers.
- Black-box solver performance visualizer (BBOviz). Developed by Zhang, Y. and Sahinidis, N.V.
- Code releases: BARON 25.2.1.
2024
- Dec New publication:
- Zhang, Y., Sahinidis, N.V. Solving continuous and discrete nonlinear programs with BARON. Comput Optim Appl (2024).
- Since the release of BARON 24.12.8, this work has become the top recommended citation for optimization scientists publishing work using the global solver BARON.
- Zhang, Y., Sahinidis, N.V. Solving continuous and discrete nonlinear programs with BARON. Comput Optim Appl (2024).
- Jun BARON Tutorial: EUROPT 2024.
- Invited talks: EUROPT 2024, EURO 2024, Panos-70.
- Code releases: BARON 24.1.30, BARON 24.5.8.
2023
- Jun Award: 2023 iSoGo Best Theory Paper Award by International Society of Global Optimization (iSoGO).
- Invited talks: WCGO 2023, PanOptiC 2023.
- Code releases: BARON 23.1.5, BARON 23.4.28, BARON 23.6.15, BARON 23.11.13.