Thanks a lot
Raf
I don't think there are any technical reports on
orthogonal representations because "orthogonal"
depends on what it is that you are trying to
represent. An orthogonal representation means
that each piece of data represents an "independent"
variable. Suppose you have two pieces of data, e.g.,
width and height. If in your particular application,
the width is "dependent" on the height, then data
fields specifying both width and height in the
genetic algorithm results in redundant information.
And because redundant information requires more "bits" of
genetic code, it will often take longer to
evolve to an optimal solution.
In other applications, the width and height
might be independent variables (orthogonal)
and you will need genes for both width and
height.
Make sense?
-George