In 2020, we conducted a study to get some clear data for comparing the performance and readability of FICO® Xpress Mosel and Python for selected (optimization) modelling tasks, publishing the results in the blog post: 4 Main Takeaways from Comparing Xpress Mosel and Python for Optimization Models.
Due to several requests, the Mosel and Python files used for that study have been updated to reflect the most recent advancements in Xpress Mosel and the Xpress Python API, respectively. Both the source and data files were added to the FICO Xpress Optimization Examples Repository:We invite you to download the files from the Xpress Examples Repository and carry out your own experiments.
Here is a summary of the key findings which have not changed materially. Yet, all interfaces have significantly improved performance compared to 2020:
Stay tuned to learn more about the newest features of FICO Xpress Solver. Until then, you are welcome to try the Community Edition and experiment with the included license (limits on problem size apply).
Due to several requests, the Mosel and Python files used for that study have been updated to reflect the most recent advancements in Xpress Mosel and the Xpress Python API, respectively. Both the source and data files were added to the FICO Xpress Optimization Examples Repository:We invite you to download the files from the Xpress Examples Repository and carry out your own experiments.
Here is a summary of the key findings which have not changed materially. Yet, all interfaces have significantly improved performance compared to 2020:
- Performance-tuned Python is up to 10x faster than Mosel at data processing.
- Mosel is up to 10x faster than Python for optimization modeling.
- Mosel performs well out-of-the-box while being closest to the natural formulation of optimization models, which supports readability and maintainability of the code.
Stay tuned to learn more about the newest features of FICO Xpress Solver. Until then, you are welcome to try the Community Edition and experiment with the included license (limits on problem size apply).
Updated 1 year ago
Version 3.0