GAMS participated in YAEM 2025 in Ankara, one of Turkey’s major operations research and industrial engineering conferences. Our team was on the ground to present talks, join a panel discussion, and meet with researchers, students, and professionals across academia and industry.
We contributed two technical presentations, “GAMS Engine SaaS: A Cloud-Based Solution for Large-Scale Optimization Problems” by Merve Demirci and “Embedding Trained Neural Networks in GAMSPy” by Hamdi Burak Usul. Additionally we provided a 45-minute expert panel. Together, these sessions drew strong interest from both academic and industry participants, leading to vibrant dialogue on Python integration, real-world applications, and the next generation of optimization tools.
Our booth became a vibrant meeting point, drawing interest from students, professors, and industry representatives.
Visitors were excited to learn about our free academic licenses, student project opportunities, and how GAMSPy bridges the gap between Python and GAMS. Interest was especially strong around Engine’s potential in university-industry collaborations and MIRO’s ability to bring optimization models to a wider, non-technical audience.
We’re excited about the connections made in Ankara and look forward to deepening our presence and collaborations in Turkey.
Embedding Trained Neural Networks in GAMSPy
By Hamdi Burak Usul
GAMSPy is a powerful mathematical optimization package that combines Python’s flexibility with GAMS’s modeling performance. GAMSPy enables previously challenging applications in the area of combining machine learning (ML) and mathematical modeling. To support these ML applications, our work introduces essential ML operations into GAMSPy, such as matrix multiplication, transposition, and norm calculations. Building on this foundation, we introduce GAMSPy “formulations”, a straightforward way to model common neural network constructs like linear (dense) layers, convolutional layers, and activation functions (ReLU, tanh and so on). When there are many good ways to formulate a construct, we implement many of them and to let user decide depending on their use case. In addition to neural network construct, we also introduce formulations for classical ML constructs such as regression trees and so on. In this talk, we demonstrate these enhancements by generating adversarial images for the German Traffic Sign Recognition Benchmark (GTSRB) using GAMSPy. We selected GTSRB because it requires a neural network that is significantly larger than many other toy examples neural networks like the ones trained for MNIST.
GAMS Engine SaaS: A Cloud-Based Solution for Large-Scale Optimization Problems
By Merve Demirci
GAMS Engine SaaS is a cloud-based service that allows users to run GAMS jobs on a scalable and flexible infrastructure, currently provided by Amazon Web Services (AWS). It was launched in early 2022 and has since attracted a variety of customers who benefit from its features, such as horizontal auto-scaling, instance sizing, zero maintenance, and simplified license handling. GAMS Engine SaaS is especially suitable for workloads that require large amounts of compute power and can be adapted to many different scenarios. In this presentation, we show a case study of a large international consultant agency that uses GAMS Engine SaaS to run Monte-Carlo simulations of a large energy system model in response to varying climate change scenarios. We describe how they leverage the GAMS Engine API to submit and monitor their jobs, how they select the appropriate instance type for each job, and how they can use custom non-GAMS code on Engine SaaS. We also discuss the challenges and benefits of using GAMS Engine SaaS for this type of application, and provide some insights into the future development of the service.