Inquiry About Sequential Evaluation and Stopping Criteria in HyperStudy

LEE_RQ

I am using HyperStudy for a multi-objective optimization project where each objective is calculated independently. Currently, my process is as follows:

1. I have set up three objectives (Objective 1, Objective 2, and Objective 3). Each objective is evaluated through separate solver calls.
2. HyperStudy first runs the solver for Objective 1, calculates the result, and then proceeds to evaluate Objective 2 and Objective 3 sequentially.
3. After all objectives are calculated, I evaluate whether the results meet the constraints.

I would like to streamline this process such that:

• After Objective 1 is calculated, the system immediately checks if its result satisfies its constraints.
• If Objective 1 does not meet the criteria, the system stops further calculations for this parameter set and moves to the next iteration.
• If Objective 1 is satisfied, the system proceeds to Objective 2, evaluates it, and applies the same logic, and so on.

Could you please advise if HyperStudy supports this kind of conditional stopping and sequential evaluation, and how to configure it?

Garima_Singh

Hi @LEE_RQ,

There is a feature of 'Model Conditions' in HyperStudy 2024 that can be tested for this requirement.

I infer the following as the requirement:

• If Objective 1 is met, Solver 2 should be invoked to evaluate Objective 2.
• Else, If Objective 1 is not met, Solver 2 should not be invoked.

You can test the 'File Exists' option such that if the a file exists in the m_1 folder (indicating meeting of Objective 1), only then Solver 2 would get invoked.

Note: you will have to write a script that checks Solver 1 response & the particular objective value. The script then writes a file (.txt/ASCII) which can be added as a Condition for Model 2 (Solver 2) to get executed.

Feature of 'Model Conditions' in HyperStudy

Many thanks.

Kind Regards

Garima Singh