Question:
Indefinite integrals of piecewise functions and their constants of integration?
?
2016-10-20 06:13:32 UTC
For the indefinite integral of a piecewise function, would each section of the piecewise function, when integrated, have its own unique constant of integration?
Three answers:
Morewood
2016-10-23 10:56:37 UTC
Short answer "NO". For ANY function f(x) defined on some (connected) interval a


Standard Proof (which you should have seen before):



Suppose first that some anti-derivative exists. That is that some F(x) is defined for x∈(a,b) such that F'(x)=f(x) for all x∈(a,b).



Now suppose that there is another anti-derivative, G(x). Let K(x) = G(x)-F(x), so that K'(x) = G'(x)-F'(x) = f(x)-f(x) = 0. Then for any x,c∈(a,b), by the mean value theorem, K(x)-K(c)=K'(z)(x-c) for some "z" between "x" and "c". But K'(z)=0, hence K(x)=K(c) for all x∈(a,b). That is, K(x) is constant. Every anti-derivative of f(x) is of the form F(x)+K for some constant K. Furthermore, it is easy checked that every F(x)+K is an anti-derivative of f(x).



QED



Notice that this theorem requires a connected domain. If the domain of f(x) is not connected, then anti-derivatives can have differ by different constants on each connected component. (But only one constant within each component, by the above theorem.) For example: f(x)=1/x² has anti-derivatives:

    F(x) = A - 1/x, for x>0; and F(x) = B - 1/x, for x<0

        for any constants A and B.

Also note that, in this case, F(x) can never be extended to a function continuous at x=0 regardless of the choices for "A" and "B". Of course, being differentiable by definition, anti-derivatives are always continuous on their entire domain.



Also notice that the theorem requires that f(x) have at least one anti-derivative defined on the entire domain, otherwise not anti-derivative means no constant of integration! (You ought to have seen examples of functions which do not have anti-derivatives.)



Normally, piecewise functions are defined on a connected domain, and so allow at most one constant of integration.



For example, if f(x) =

    2x+1   for x<1

      3     for 1≤x≤5

    3x²-72   for x>5

Then one anti-derivative is: F(x) =

    x² + x   for x<1

    3x - 1   for 1≤x≤5

    x³-72x+249   for x>5

Notice the "-1" and the "249" required to keep F(x) continuous, hence differentiable, at x=1 and x=5. Now then, EVERY anti-derivative of f(x) must take the form:

    x² + x + k   for x<1

    3x - 1 + k   for 1≤x≤5

    x³-72x+249+k   for x>5

For any constant "k", the ONE constant of integration.



Notice that, in the example to which Cidyah refers, the original piecewise function is undefined at both 0 and 20 (disconnected domain with 3 sections, hence 3 constants) and all the anti-derivatives have derivatives at neither 0 nor 20.
J
2016-10-20 07:45:21 UTC
You should expect the indefinite integral to be a piecewise defined function. As such each piece should have its own constant of integration.



However, if you want the indefinite integral to be continuous, then once you pick the constant of integration for one piece, which will be an arbitrary constant C, then that will determine the constant for each of the other pieces, each of which will be expressed in terms of C.
cidyah
2016-10-20 06:32:06 UTC
Yes. Note C1, C2, and C3 in the link:

https://www.youtube.com/watch?v=gzMYxHTXHqA


This content was originally posted on Y! Answers, a Q&A website that shut down in 2021.
Loading...