Memory overflow and unexpected shutdown during Constrained Bayesian optimization with multiple objectives

[BinglinW]

Describe the bug
Hello, I am a researcher using Bayesian optimization for material preparation. I wrote a Bayesian optimization with constraints, where the optimization is a mixing constraint: x1+x2+... +xn< 1.
When the sample size is small (less than 40), the program works fine. Occurs when the sample size is greater than 40, TensorFlow ran out of memory by 10%, which I had solved earlier by switching to a better computer. However, my current sample size is 55. Not only did TensorFlow run out of memory by 10%, but the computer would restart with blue screen.

I would like to know how to fix this problem, for example if it is possible to limit memory usage. Data can be sent privately if needed.

System information

• OS: win 10
• Python version: 3.7 and 3.9
• Trieste version: 0.13.3
• TensorFlow version: 2.10.1
• GPflow version: 2.6.4

[uri-granta]

Thanks for reporting this and sorry that it was not noticed for so long! Are you still trying to use trieste or have you given up?