Structural bias in population-based algorithms
- Submitting institution
-
Heriot-Watt University
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 10596362
- Type
- D - Journal article
- DOI
-
10.1016/j.ins.2014.11.035
- Title of journal
- Information Sciences
- Article number
- -
- First page
- 468
- Volume
- 298
- Issue
- -
- ISSN
- 0020-0255
- Open access status
- Out of scope for open access requirements
- Month of publication
- March
- Year of publication
- 2015
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
4
- Research group(s)
-
-
- Citation count
- 30
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Structural bias (SB) is a fundamental but ill-understood attribute of all optimization algorithms (broadly analogous to inductive bias in machine learning). For the first time, a mathematical result is derived concerning concerning SB in population-based stochastic optimization, revealing previously unknown consequences of SB which impact on all optimization practitioners and researchers, especially in terms of designing settings and operators (i) increasing population size will magnify inherent SB, potentially misleading the search; (ii) effects of SB are magnified by landscape complexity. Empirical validation is comprehensive and careful (ensuring we eliminate artefacts of the random number generator).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
