Given a function $f(x)$, its inverse $g(x)$ is defined as $g(f(x)) \equiv x$.
In light of this, the n-th derivative of $g(x)$ can be recursively calculated as follows:
list = {};
For[n = 1, n <= 6, n++,
eqn = D[g[f[x]], {x, n}] == D[x, {x, n}];
var = Derivative[n][g][f[x]];
If[n == 1,
list = Join[list, Solve[eqn, var][[1]]],
list = Join[list, (Solve[eqn, var] /. list)[[1]]]
]
]
list // Expand // TableForm
In MMA you can get this very simply by writing something like this:
Derivative[n][InverseFunction[f]][f[x]]
The question that arises is the following: is the latter formulation using a recursive process like the one shown above or is there a non-recursive formulation?
Answer
D[InverseFunction[f][x], {x, 6}]
Towards the question whether this is computed recursively: I guess so from analyzing the Trace
produced by executing the code:
Table[
Length@Trace[
D[InverseFunction[f][x], {x, k}]
],
{k, 1, 12}]
{6, 7, 7, 10, 13, 18, 25, 36, 51, 73, 103, 145}
Comments
Post a Comment