MeesterDaan (talk | contribs) (→Misc) |
MeesterDaan (talk | contribs) (→Misc) |
||
Line 114: | Line 114: | ||
− | * We got simulated annealing to work in a [http://heuristieken.nl/resources/(2020)Dahmanietal-PFPSimulatedAnnealing paper published at ICCC'20 | + | * We got simulated annealing to work in a [http://heuristieken.nl/resources/(2020)Dahmanietal-PFPSimulatedAnnealing.pdf paper] published at ICCC'20. |
Revision as of 20:05, 8 September 2020
This page is dedicated to the EVOSTAR'19 publication "Paintings, Polygons and Plant Propagation".
Paintings
These are the target bitmaps used in our experimental runs.
Mona Lisa | Mondriaan | Klimt | Bach |
---|---|---|---|
Dali | Jackson Pollock | Starry Night |
---|---|---|
Painting provenance (recovered 31-9-2018):
BACH: Wikipedia, Public Domain (https://en.wikipedia.org/wiki/Johann_Sebastian_Bach)
DALI: Wikiart, Fair Use (https://www.wikiart.org/en/salvador-dali/the-persistence-of-memory-1931)
MONA LISA: https://www.wikiart.org/en/leonardo-da-vinci/mona-lisa
MONDRIAAN: https://www.wikiart.org/en/piet-mondrian/composition-with-red-blue-and-yellow-1930
POLLOCK: https://www.wikiart.org/en/jackson-pollock/convergence-1952
STARRY NIGHT: https://www.wikiart.org/en/vincent-van-gogh/the-starry-night-1889
KLIMT: https://www.wikiart.org/en/gustav-klimt/the-kiss-1908
Get all the painting bitmaps in one zip file.
PolygonConstellation
These are the constellations from our last figure:
New Paintings
Salvator | Lady Ermine |
---|---|
Resources & Media
A movie clip of the algorithm at work.
- [graph typicalruns][graph best end results][link to data records] ... available on request.
- Here's our source code.
Errata
- In equation (7), the factor (1−fi) should be just (fi). The typo is just in the paper, not in the source code, and therefore doesn't affect the results (see organisms.py, line 138).
- In equation (9), the factor 9v/4 is incorrect; the experiments were done with 13v/5. The factor 1/nmax from the same formula was never in the experiment (see organisms.py, line 143). It is most likely a uncorrected textual remnant from an earlier probe. In general, we have been looking for a good way to transfer the PPA-algorithm to this problem, and many adaptations are possible.
Misc
- A simplified version of the problem was proven to be NP-hard in 2020.
- We got simulated annealing to work in a paper published at ICCC'20.