Abstract
There exist general transforms that convert pseudo-Boolean functions into k-bounded pseudo-Boolean functions, for all k ≥ 2. In addition to these general transforms, there can also exist specialized transforms that can be applied in special cases. New results are presented examining what happens to the "bit flip" neighborhood when transforms are applied. Transforms condense variables in a particular order. We show that different variable orderings produce different results in terms of problem difficulty. We also prove new results about the embedding of the original function in the new k-bounded function. Finally, this paper also looks at how parameter optimization problems can be expressed as high precision k-bounded pseudo-Boolean functions. This paper lays a foundation for the wider application of evolutionary algorithms to k-bounded pseudo-Boolean functions.
Original language | English (US) |
---|---|
Title of host publication | GECCO 2020 - Proceedings of the 2020 Genetic and Evolutionary Computation Conference |
Publisher | Association for Computing Machinery |
Pages | 760-768 |
Number of pages | 9 |
ISBN (Electronic) | 9781450371285 |
DOIs | |
State | Published - Jun 25 2020 |
Event | 2020 Genetic and Evolutionary Computation Conference, GECCO 2020 - Cancun, Mexico Duration: Jul 8 2020 → Jul 12 2020 |
Publication series
Name | GECCO 2020 - Proceedings of the 2020 Genetic and Evolutionary Computation Conference |
---|
Conference
Conference | 2020 Genetic and Evolutionary Computation Conference, GECCO 2020 |
---|---|
Country/Territory | Mexico |
City | Cancun |
Period | 7/8/20 → 7/12/20 |
Bibliographical note
Publisher Copyright:© 2020 ACM.
Keywords
- Bit representations
- Combinatorial optimization
- Epistatsis
- Pseudo-boolean functions