Optimization, zoo and deepseek

A few years ago, I wrote Optimization : simply do more with less, zoo, buses and kids in order to explain what optimization is and how IBM CPLEX can help.

And 2 years ago when ChatGPT OpenAI was born I ould not help trying to generate the model with this new tool.

This worked fine and I tried the same within #watsonx IBM with other LLMs like #Llama Meta or even with #watsonx code assistant last month.

Of course, to be fair in late January 2025 I had to try with DeepSeek AI !

And it works quite well!

First, let' ask this new LLM:

Then we get

And

which leads to

And then we get the result

Congratulations DeepSeek . Will this open optimization to many more use cases ? At least starting with a small optimization model is not limited any longer to a happy few. OPL and algebraic modeling languages made optimization much easier to use and I hope generative AI will also help. I look forward to your comments.

NB: You can run #CPLEX and #DeepSeek in #watsonx #AI and also without any #watsonx