New adaptive interior point algorithms for linear optimization
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
Interior Point Methods (IPMs) have shown their power in solving large scale optimization problems. In this thesis, using the notion of self-regularity, various classes of algorithms are proposed. First, we present an adaptive technique to compute the target barrier parameter. Using this adaptive technique a single step large-update algorithm is proposed that enjoys better worst case iteration complexity than the classical large-update IPMs. A new variant of infeasible IPMs, using the same adaptive technique, is proposed that enjoys better worst case iteration complexity than the classical infeasible IPMs. By further use of the adaptive technique a new family of predictor-corrector IPMs is proposed that are different both in the predictor and the corrector steps from their classical counter parts. Worst case iteration complexity analysis reveals significant improvement for some special cases over the classical analogues. In the second part of thesis we analyze the theoretical behavior of the most widely used algorithm in IPM based software packages, the so called Mehrotra-type predictor-corrector algorithm. This analysis reveals some drawbacks of the algorithm that motivated us to modify it. We combine it with a safeguard that prevents the drawback we observed and enables us to prove strong theoretical results about the new algorithm. Motivated by this drawback, we analyze it from different perspective. In the new approach, contrary to the existing variants in the literature, first we fix the step size and then aim to compute the smallest value of the barrier parameter that ensures the prescribed step size is feasible. After the barrier parameter is computed, the algorithm computes the next iterate by a line-search as usual. The worst case behavior of the proposed algorithms is discussed. Finally, some limited encouraging computational results are reported.
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 | 0.000 |
| 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