Here's a worthy computational aerodynamic project that occured
to me today. I think it would not be too difficult for a young
student to execute.
Purpose.
Develop an airfoil section superior in some measure than
any previously developed.
Method.
Obtain an airfoil evaluation program.
(I saw a program on offer yesterday for $95 in sci.engr to which
airfoil coordinates are input and which plots polar curves.)
Obtain a copy of the Byte article *(from spring of this year) which
discusses a very compact method of optimising a set of parameters
(the 'genome') according to a given equation describing the quality
criterion which relates the parameters, using the genetic algorithm.
(* title: "Differential Evolution")
Create a simple pair of polynomial equations whose beginning and
end-points touch along the x-axis as a compact way of describing a
novel airfoil.
Write some code that can pass parameters to the polynomial equations.
Write some code that can pass the coordinates from the polynomial
equations to the airfoil evaluation program.
Write some code that can express the polar plots as a numeric
quality criterion.
Write some code that passes the value of the criterion back to the
genetic algorithm so that it will continue to select superior
candidates for survival and reproduction.
I am supposing this might amount to 50 lines of code or less.
I reckon that with perhaps six parameters describing an upper surface,
and a similar number describing the lower surface a genetic algorithm
could hardly fail to produce a superior design within a day or
two of PC run time.