Diploidy in evolutionary algorithms for dynamic optimization problems
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
Purpose – In this work, the authors show the performance of the proposed diploid scheme (a representation where each individual contains two genotypes) with respect to two dynamic optimization problems, while addressing drawbacks the authors have identified in previous works which compare diploid evolutionary algorithms (EAs) to standard EAs. The paper aims to discuss this issue. Design/methodology/approach – In the proposed diploid representation of EA, each individual possesses two copies of the genotype. In order to convert this pair of genotypes to a single phenotype, each genotype is individually evaluated in relation to the fitness function and the best genotype is presented as the phenotype. In order to provide a fair and objective comparison, the authors make sure to compare populations which contain the same amount of genetic information, where the only difference is the arrangement and interpretation of the information. The two representations are compared using two shifting fitness functions which change at regular intervals to displace the global optimum to a new position. Findings – For small fitness landscapes the haploid (standard) and diploid algorithms perform comparably and are able to find the global optimum very quickly. However, as the search space increases, rediscovering the global optimum becomes more difficult and the diploid algorithm outperforms the haploid algorithm with respect to how fast it relocates the new optimum. Since both algorithms use the same amount of genetic information, it is only fair to conclude it is the unique arrangement of the diploid algorithm that allows it to explore the search space better. Originality/value – The diploid representation presented here is novel in that instead of adopting a dominance scheme for each allele (value) in the vector of values that is the genotype, dominance is adopted across the entire genotype in relation to its homologue. As a result, this representation can be extended across any alphabet, for any optimization function.
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 | 0.001 | 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 | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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