Why Periodicity Helps with Integration
A periodic function repeats its values at regular intervals. That repetition can make a definite integral easier to evaluate: a long interval may contain several copies of the same full-period contribution, and an interval can often be translated to one with more convenient endpoints.
The key point is that the integral over a full period does not depend on where that period begins. This is true even if the starting point is not a familiar endpoint such as \(0\). Once this fact is established, an interval spanning several periods can be separated into complete periods and a shorter remainder.
Periods and Periodic Functions
For example, \(3+\sin x\) has period \(2\pi\), because \(\sin(x+2\pi)=\sin x\). Its values repeat, but the function is not identically zero. Periodicity therefore does not by itself determine the value of an integral; we also need the integral over one period.
In what follows, assume that \(f\) is continuous, or at least integrable on every finite interval under consideration, and is periodic on \(\mathbb{R}\). These conditions ensure the definite integrals used below exist. Write
for the integral over one period, initially using any starting point \(c\). We will show that this value is independent of \(c\).
Every Full Period Has the Same Integral
Here is a proof using additivity of definite integrals and a translation of the integration variable. First consider a shift \(s\) with \(0\leq s<T\). Split the shifted interval at \(a+T\):
In the second integral, set \(u=x-T\). Then \(dx=du\), and its endpoints \(a+T\) and \(a+s+T\) become \(a\) and \(a+s\). Periodicity gives \(f(u+T)=f(u)\), so
Adding this to the first part recombines adjacent intervals:
For a general shift \(h\), write \(h=mT+s\), where \(m\) is an integer and \(0\leq s<T\). Translation by \(mT\) preserves the function values, since repeated use of \(f(x+T)=f(x)\) gives \(f(x+mT)=f(x)\) for every integer \(m\). Thus shifting both endpoints by \(mT\) leaves the integral unchanged. The result for the remaining shift \(s\) follows from the calculation above. This proves the theorem for every \(h\).
In particular, choosing \(a=0\) shows that every full-period integral equals \(\int_0^T f(x)\,dx\). Denote this common value by \(P\). The value \(P\) can be positive, negative, or zero, depending on \(f\).
Translating an Integral by a Whole Number of Periods
A related shortcut applies when both endpoints are shifted by the same integer multiple of \(T\). For any integer \(k\), substitution \(u=x-kT\) gives
This identity lets us move an interval to a more convenient location without changing its integral. It does not say that an arbitrary interval has the same integral as a full period: the interval’s length matters. In particular, an interval of length \(T\) has integral \(P\), while an interval of a different length generally includes a different amount of the repeated pattern.
Worked Example: Translating a Full Period of a Trigonometric Function
Evaluate \(\int_{\pi/3}^{7\pi/3}(2+\cos x)\,dx\). The function \(2+\cos x\) has period \(2\pi\), and the interval has length \(7\pi/3-\pi/3=2\pi\). Therefore the integral is one full-period integral, equal to the integral over \([0,2\pi]\).
Check by evaluating an antiderivative at the original endpoints. Since \(2x+\sin x\) is an antiderivative,
The endpoint sine terms cancel, confirming the translated full-period calculation.
Several Full Periods and a Remainder
Suppose \(b\geq a\). Divide the interval length by \(T\): choose an integer \(q\geq 0\) and a remainder \(r\) with \(0\leq r<T\) such that \(b-a=qT+r\). Split the interval into \(q\) consecutive pieces of length \(T\) and one final piece of length \(r\). Each full-period piece contributes \(P\), so
This formula separates the repeated contribution from the part that does not make up a complete period. The remainder must still be evaluated; periodicity alone does not make it vanish.
Worked Example: Two Full Periods and a Partial Period
Let \(f\) be the triangular periodic function of period \(2\) defined on one period by
Evaluate \(\int_{-1/2}^{9/2}f(x)\,dx\). The interval length is \(9/2-(-1/2)=5=2(2)+1\), so it contains two full periods and a remainder of length \(1\). The full-period integral is the area of two triangles, each with base \(1\) and height \(1\):
The remainder can be translated by \(2\) to \([3/2,5/2]\), or by periodicity evaluated as the interval \([-1/2,1/2]\). Split that interval at \(0\). On \([-1/2,0]\), periodicity gives \(f(x)=f(x+2)=2-(x+2)=-x\); on \([0,1/2]\), \(f(x)=x\). Thus
Therefore the two full periods contribute \(2P=2\), and the remainder contributes \(1/4\):
Check using an antiderivative. Define \(G(x)=x^2/2\) on \([0,1]\), and \(G(x)=2x-x^2/2-1\) on \([1,2]\), then extend it by \(G(x+2)=G(x)+1\). On the first segment \(G'(x)=x=f(x)\); on the second \(G'(x)=2-x=f(x)\). The formulas and derivatives agree at the join, and the extension has derivative \(f\) because \(f\) is periodic. Now \(G(1/2)=1/8\), so \(G(-1/2)=G(3/2)-1=7/8-1=-1/8\). Also \(G(4+1/2)=G(1/2)+2=17/8\). Hence
which confirms the result.
Using the Shortcut When the Remainder Is Zero
If an interval has length exactly \(nT\) for a positive integer \(n\), it consists of \(n\) full periods and has integral \(nP\), regardless of its starting point. The function need not be positive, and the result need not be zero. The value depends on the net signed contribution over one period.
Worked Example: Four Full Periods with a Nonzero Integral
Evaluate \(\int_{-\pi/2}^{15\pi/2}(3+\sin x)\,dx\). The function has period \(2\pi\). The interval length is
so it consists of four full periods. Over one period,
Consequently, the integral is \(4P=24\pi\). Check directly with the same antiderivative \(3x-\cos x\):
Both cosine terms are zero, so the direct evaluation agrees with the four-period calculation.
A Common Mistake: Assuming Periodic Integrals Vanish
A common error is to assume that the integral over a period must be zero because the function repeats. Repetition says that the signed area repeats; it does not say that positive and negative contributions cancel. For \(f(x)=2+\sin x\), the incorrect calculation would be
perhaps from noticing that the sine term has zero integral over a period. But the constant term contributes:
The correction is to calculate \(P\) for the whole function, including every term, before multiplying by the number of full periods. Periodicity gives a shortcut for repeating that value; it does not replace the calculation of the value itself.
When This Shortcut Matters
Periodicity is useful when endpoints are far apart, shifted away from convenient values, or arranged so the interval contains several full cycles. Instead of evaluating a long integral in one step, identify a period, compute one full-period integral, count the complete repetitions, and then handle any leftover interval.
Verify that \(f(x+T)=f(x)\) for all relevant \(x\), with \(T>0\).
Find \(P=\int_0^T f(x)\,dx\), or use any more convenient interval of length \(T\).
Write the interval length as \(qT+r\), where \(0\leq r<T\).
Use periodicity to move the remainder to convenient endpoints, then add its integral to \(qP\).
For a reversed interval, the usual orientation rule applies: \(\int_b^a f(x)\,dx=-\int_a^b f(x)\,dx\). The periodicity calculations above can be applied to the interval in increasing order first, with the sign reversed afterward.
Check Your Understanding
Use the period and the integral over one period to justify each answer; do not assume that the integral of a periodic function over one period is zero.
- If \(f\) has period \(T\), why is \(\int_a^{a+T}f(x)\,dx\) independent of \(a\)?
- A function has period \(3\) and \(\int_0^3 f(x)\,dx=5\). What is \(\int_{7}^{16} f(x)\,dx\)?
- For a periodic function with period \(T\), what does translating both bounds by \(4T\) do to a definite integral?
- An interval has length \(5T/2\). How many complete periods does it contain, and what is the length of the remainder?
- Give a reason that periodicity alone does not imply \(\int_0^T f(x)\,dx=0\).
