1
$\begingroup$

When you try to use the Multivariable option of Series, it doesn't work as intended, since it keeps terms of higher order than intended.

Example: Expand a two-variable gaussian exponential to second order.

In[1]:= f[x_, y_] := Exp[-x^2 - y^2]
In[2]:= Normal[Series[f[x, y], {x, 0, 2}, {y, 0, 2}]]
Out[2]= 1 - y^2 + x^2 (-1 + y^2)

Here the last term is on fourth order. I found a solution to this here Multivariable Taylor expansion does not work as expected.

Applying this to my problem yields this,

In[3]:= Normal[Series[f[x t, y t], {t, 0, 2}]] /. t -> 1
Out[3]= 1 - x^2 - y^2

Which is the correct expansion to second order.

Now I intend to create a function to apply all the process in one step,

In[4]:=MultiSeries[exp_] := Module[{f},
f[x_, y_] := Evaluate[exp];
Normal[Series[f[x t, y t], {t, 0, 2}]] /. t -> 1
]

The problem is that by doing that, it yields this,

In[5]:= MultiSeries[Exp[-x^2 - y^2]]
Out[5]= E^(-x^2 - y^2)

The problem doesn't seem to be on the compound expression since,

In[6]:= Clear[f]; f[x_, y_] := Exp[-x^2 - y^2]; Normal[Series[f[x t, y t], {t, 0, 2}]] /. t -> 1

Yields,

Out[6]= 1 - x^2 - y^2

This leads me to believe that the problem is in the Module function. Is there any way to solve this?

$\endgroup$
  • 4
    $\begingroup$ It's a scoping issue involving the dummy pattern variables not being recognized as the "same" as the ones used in the input. A way to avoid this is to make the variables explicit in the Module input. Could do multiSeries[exp_, vars_] := Module[{t}, Normal[Series[exp /. Thread[vars -> t*vars], {t, 0, 2}]] /. t -> 1] and try it on multiSeries[Exp[-x^2 - y^2], {x, y}]. $\endgroup$ – Daniel Lichtblau Nov 21 '18 at 15:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.