Re: Confusing Result with Series

*To*: mathgroup at smc.vnet.net*Subject*: [mg121268] Re: Confusing Result with Series*From*: Andrzej Kozlowski <akoz at mimuw.edu.pl>*Date*: Tue, 6 Sep 2011 04:00:08 -0400 (EDT)*Delivered-to*: l-mathgroup@mail-archive0.wolfram.com*References*: <201109051105.HAA28332@smc.vnet.net>

Note that these are Puiseaux series. I think in the case of such series the "truncation order" in expression of the form Series[f[x], {x, 0, n}] is not defined uniquely but depends on how the expansion is done. This is actually explained in the tutorial "MakingPowerSeriesExpansions": The procedure that Series follows in constructing a power series is largely analogous to the procedure that N follows in constructing a real-number approximation. Both functions effectively start by replacing the smallest pieces of your expression by finite-order, or finite-precision, approximations, and then evaluating the resulting expression. If there are, for example, cancellations, this procedure may give a final result whose order or precision is less than the order or precision that you originally asked for. Like N, however, Series has some ability to retry its computations so as to get results to the order you ask for. In cases where it does not succeed, you can usually still get results to a particular order by asking for a higher order than you need. However, there is another way to make sure you get a series with the specified truncation term. Just use this form f[x]+ O[x]^a : In[98]:= x/Sqrt[1+Sqrt[x]]+O[x]^(3/2) Out[98]= x+O(x^(3/2)) In[99]:= Sqrt[x^2/(1+Sqrt[x])]+O[x]^(3/2) Out[99]= x+O(x^(3/2)) In[100]:= x/Sqrt[1+Sqrt[x]]+O[x]^2 Out[100]= x-x^(3/2)/2+O(x^2) In[101]:= Sqrt[x^2/(1+Sqrt[x])]+O[x]^2 Out[101]= x-x^(3/2)/2+O(x^2) Andrzej Kozlowski On 5 Sep 2011, at 13:05, jschwab wrote: > Hi Mathematica Gurus, > > I am seeing confusing behavior with the Series command, using > Mathematica 7.0.1 on Mac OS X. > Specifically, it is not always truncating at the order I would expect. > Here's a a simple example of the problem I'm having. > > I expect that the result of a a series expansion of > \frac{x}{\sqrt{1 + \sqrt{x}}} > to first order will be > x + O(x^{3/2}) > no matter how I write the input. > > Instead, I see the following behavior. > > In[175]:= Series[x/Sqrt[1 + Sqrt[x]], {x, 0, 1}] > Out[175]= SeriesData[x, 0, {1, Rational[-1, 2]}, 2, 4, 2] > > In[176]:= Series[Sqrt[x^2/(1 + Sqrt[x])], {x, 0, 1}] > Out[176]= SeriesData[x, 0, {1}, 2, 3, 2] > > Is this a known issue? Or some sort of expected behavior that I > haven't understood? > A naive Google search didn't reveal anything of particular relevance. > > Thanks, > Josiah >

**References**:**Confusing Result with Series***From:*jschwab <jschwab@gmail.com>