A Chebyshev series is stored using the GSL::Cheb class.
GSL::Cheb.new(n)
GSL::Cheb.alloc(n)
-
This create an instance of the GSL::Cheb class for a Chebyshev series of order n.
GSL::Cheb#init(f, a, b)
-
This computes the Chebyshev approximation the function f over the range (a,b) to the previously specified order. Where f is a GSL::Function object. The computation of the Chebyshev approximation is an O(n^2) process, and requires n function evaluations.
ex: Approximate a step function defined in (0, 1) by a Chebyshev series of order 40.
f = GSL::Function.new { |x|
if x < 0.5
0.25
else
0.75
end
}
cs = GSL::Cheb.new(40)
cs.init(f, 0, 1)
GSL::Cheb#eval(x)
-
This evaluates the Chebyshev series at a given point x.
GSL::Cheb#eval_n(n, x)
-
This evaluates the Chebyshev series at a given point x, to (at most) the given order n.
GSL::Cheb#calc_deriv()
GSL::Cheb#deriv()
-
This computes the derivative of the series, and returns a new GSL::Cheb object which contains the computed derivative. The reciever is not changed.
GSL::Cheb#calc_integ()
GSL::Cheb#integ()
-
This computes the integral of the series, and returns a new GSL::Cheb object which contains the computed integral coefficients. The reciever is not changed.
#!/usr/bin/env ruby
require("gsl")
f = GSL::Function.new { |x|
if x < 0.5
0.25
else
0.75
end
}
n = 1000
order = 40
cs = GSL::Cheb.new(order)
cs.init(f, 0, 1)
x = Vector.linspace(0, 1, n)
ff = f.eval(x)
r10 = cs.eval_n(10, x)
r40 = cs.eval(x)
GSL::graph(x, ff, r10, r40)
See also the example scripts in samples/cheb/
.
prev
next
Reference index
top