DUQO: Design Uncertainty Quantification and Optimization framework

The repo is under ongoing development. Breaking changes may occur in next versions.

Current stable version can be installed using pip pip install duqo

Alternatively, clone the master branch and use pip install -e .

See examples for example usage. Documentation is on its way.

Reliability Analysis and Reliability-based Robust Design Optimization

Generally, given one or more limit state functions of form

as well as the input distributions

as parametrized by , uncertainty quantification, i.e. reliability-analysis, seeks to compute the probability of failure

To solve the uncertainty optimization, i.e. reliability-based robust design optimization, problem, the evaluation of as well as the expectations and variances of the objective functions

and possible deterministic constraints

with respect to the distribution parameters is required. Besides the input distributions, duqo takes the objectives , the limit states and the constraints as input and wraps them with problem.obj_con to be used by a generic gradient-free optimization algorithm.

Citation

If this repo helped you, I would appreciate citations:

C. Bogoclu, T. Nestorović, D. Roos; Local Latin Hypercube Refinement for Multi-objective Design Uncertainty Optimization, Applied Soft Computing (2021)

Contribution

Contributions welcome as there is a long road ahead to make this research code to a usable one.