Hi Everybody: I've a question about "convolution". How can I afford the convolution of different scaled functions?... In Example: F (2*X/T) in convolution with G ( X/T ) ? Is this equals to : Int ( F(2*w/T) * G((X-w)/T), w=-infinity.. infinity)? Am I wrong?... The particular problem is : Convolution of ( T*Triangle(2*X/T) ) with ( shah( X/T ) ), with Shah (x) = Sum( Dirac (x-n), n= - infinity.. infinity). Thanks for helping!

HI to all!

I've a math question about the Hessian Matrix. I know that it's used for clasify critical points of a 3-D function, but what happens if the Hessian is undefined (Hessian = 0) in a point?... How i can analyze the critical point?...

A particular sample is:  f(x,y) = (y-x)^2*(y+x).

The gradient vector of this function is: grad(f)= [ 3x^2 - 2xy - y^2 , -x^2 - 2xy + 3y^2 ]. To find critical points we have to do grad(f)=0 (vector). And in this case it's easy to verify that the critical points lie in the Y=X line.

Hi!,

I've a problem with special waveforms like Square, Triangular and SawTooth. I've to solve a RLC circuit ODE for the capacitor's charge  (q(t)) as a function of time with an Voltage input of a Square Waveform, Triangular WaveForm and a SawTooth Waveform. I solved the problem with one oscillation for each waveform, but the problem says that we have to work with the periodical characteristic function for each of them.

My question : How can I transform one oscillation in a periodical function with maple? (repeating that oscillation along the time).