This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
mind:atlas-evolutionary-investigation-1 [2016/05/02 22:21]
martindm [Something Trickier]
mind:atlas-evolutionary-investigation-1 [2016/05/02 22:27] (current)
martindm [Conclusions]
Line 32: Line 32:
 [{{mind:​ga_003.png|Alignment task}}] [{{mind:​ga_003.png|Alignment task}}]
 +== Conclusions ==
 +Using deliberate domain-specific mutations, we can achieve extremely efficient exploration of a space on simple tasks. If we can break down flow generation into a series of these simple tasks, then we can do flow generation quite cheaply.
 +I need to do experiments on a full flowMap involving global properties such as branching factor, connections,​ segment length, angle, etc. One of the tricky parts of evolutionary algorithms is selecting a fitness function that is reasonably cheap to compute. Global fitness functions tend to be quite expensive on complex domains. However, I think I can mitigate this by using global properties which can be computed on-line (that is, they don't need to be re-evaluated from scratch. Each small change can incrementally update the properties listed above).
mind/atlas-evolutionary-investigation-1.1462249308.txt.gz ยท Last modified: 2016/05/02 22:21 by martindm
Back to top
CC Attribution-Share Alike 4.0 International
chimeric.de = chi`s home Valid CSS Driven by DokuWiki do yourself a favour and use a real browser - get firefox!! Recent changes RSS feed Valid XHTML 1.0