Adaptive Constraint Handling in Optimization of Complex Structures by Using Machine Learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Optimization of complex systems requires robust and computationally efficient global search algorithms. Constraints make this a very difficult task, significantly slowing down an algorithm, and can even prevent finding the true Pareto front. This study continues the development of a recently proposed repair approach that exploits infeasible designs to increase computational efficiency of a prominent genetic algorithm, and to find a wider spread of the Pareto front. This paper proposes adaptive and automatized discovery of sensitivity of constraints to variables, i.e. the link, which needed direct designer’s input in the previous version of the repair approach. This is achieved by using machine learning in the form of artificial neural networks (ANN). A surrogate model is afterwards utilized in optimization based on ANN. The proposed approach is used for the recently proposed constraint handling implemented into NSGA-II optimization algorithm. The proposed framework is compared with two other constraint handling methods. The performance is analyzed on a structural optimization of a 178 m long chemical tanker which needs to fulfil class society’s criteria for strength. The results show that the proposed framework is competitive in terms of convergence and spread of the front. This is achieved while discovering the link automatically using ANN, without an input from a user. In addition, computational time is reduced by 60%.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it