Probabilistic Design and OPTimisation framework (PDOPT)

A python framework for set-based design space exploration without explicit elimination rules. It impements a set-based approach for mapping the requirements on the design space, using a probabilistic surrogate model trained on the provided design model. This procedure ensures to identify the best candidate areas of the design space with the minimum number of assumptions of the design of the system.

The framework process follows two steps:

• Exploration Phase. After breaking down the design space into sets (i.e. sub-spaces), the code evaluates the probability of satisfying the constraints there. Sets that have a low probability are eliminated.
• Search Phase. Surviving sets are further explored by introducing a local optimisation algorithm (GA-based), and recovering the locally optimal design points.

The final output is an aggregation of locally optimal points which constitute a rich database of design alternatives capable to minimally satisfy the requirements set by the user.

This is the archived 0.5.1 version which has been submitted to the Journal of Open Source Software (JOSS). Latest version can be found on GitHub: spinjet/pdopt-code at 0.5.1 (github.com)

This software was developed within Project FutPrint50, with EU Horizon 2020 Grant No. 875551. The authors wants to thank all the researchers in the project who contributed with their input to shape the framework.

Copyright (c) 2021 Cranfield University. This software is released under the permissive MIT License.