Let f : R + R be continuous, with
00 s_, If(x)l dx < o.
Show that there is a sequence (x,,) such that x, -+ 00, x, f (x,) 4 0, and
x, f(-xc,) 4 0 as n -+ o.

 

To show that there is a sequence (xn)(x_n) such that xn→∞x_n \to \infty, xnf(xn)→0x_n f(x_n) \to 0, and xnf(−xn)→0x_n f(-x_n) \to 0 as n→∞n \to \infty, we will leverage the fact that the integral of ∣f(x)∣|f(x)| over the entire real line is finite. Here’s a step-by-step proof:

Proof:

  1. Given Conditions:
    • f:R→Rf: \mathbb{R} \to \mathbb{R} is continuous.
    • ∫−∞∞∣f(x)∣ dx<∞\int_{-\infty}^{\infty} |f(x)| \, dx < \infty.
  2. Implications of the Integral Condition:
    • Since ∫−∞∞∣f(x)∣ dx<∞\int_{-\infty}^{\infty} |f(x)| \, dx < \infty, the integrand ∣f(x)∣|f(x)| must approach zero as x→±∞x \to \pm \infty. This follows from the fact that if ∣f(x)∣|f(x)| did not approach zero, the integral would diverge.
    • Formally, for any ϵ>0\epsilon > 0, there exists M>0M > 0 such that ∫∣x∣>M∣f(x)∣ dx<ϵ\int_{|x| > M} |f(x)| \, dx < \epsilon.
  3. Constructing the Sequence xnx_n:
    • We need to find a sequence xn→∞x_n \to \infty such that both xnf(xn)→0x_n f(x_n) \to 0 and xnf(−xn)→0x_n f(-x_n) \to 0.
  4. Using the Integral Condition to Define xnx_n:
    • Consider the function g(x)=xf(x)g(x) = x f(x). We need to show that g(x)→0g(x) \to 0 as x→±∞x \to \pm \infty.
    • Assume for contradiction that xf(x)x f(x) does not approach zero as x→∞x \to \infty. Then there exists ϵ>0\epsilon > 0 and a subsequence (xnk)(x_{n_k}) such that xnk∣f(xnk)∣≥ϵx_{n_k} |f(x_{n_k})| \geq \epsilon.
  5. Estimating the Integral on the Contradicting Subsequence:
    • Let xnkx_{n_k} be a subsequence such that xnk→∞x_{n_k} \to \infty and xnk∣f(xnk)∣≥ϵx_{n_k} |f(x_{n_k})| \geq \epsilon.
    • Consider the intervals Ik=[xnk−δ,xnk+δ]I_k = [x_{n_k} – \delta, x_{n_k} + \delta] for some small δ>0\delta > 0. The length of each interval is 2δ2\delta.
    • Since ∣f(x)∣|f(x)| is continuous, there exists δ>0\delta > 0 such that ∣f(x)∣≥ϵ2xnk|f(x)| \geq \frac{\epsilon}{2x_{n_k}} for x∈Ikx \in I_k.
  6. Calculating the Contribution to the Integral:
    • The contribution of ∣f(x)∣|f(x)| over the interval IkI_k is at least ϵ2xnk⋅2δ=ϵδxnk\frac{\epsilon}{2x_{n_k}} \cdot 2\delta = \frac{\epsilon \delta}{x_{n_k}}.
    • Summing over the subsequence IkI_k, we have: ∑k∫Ik∣f(x)∣ dx≥∑kϵδxnk.\sum_k \int_{I_k} |f(x)| \, dx \geq \sum_k \frac{\epsilon \delta}{x_{n_k}}.
    • Since xnk→∞x_{n_k} \to \infty, the terms 1xnk\frac{1}{x_{n_k}} become arbitrarily small, making the sum finite. However, this contradicts the assumption that the integral of ∣f(x)∣|f(x)| is finite unless xnkf(xnk)→0x_{n_k} f(x_{n_k}) \to 0.
  7. Conclusion:
    • Therefore, xf(x)→0x f(x) \to 0 as x→∞x \to \infty.
    • By a symmetric argument, −xf(−x)→0-x f(-x) \to 0 as x→−∞x \to -\infty.
    • Thus, we can construct the sequence xnx_n such that xn→∞x_n \to \infty, xnf(xn)→0x_n f(x_n) \to 0, and xnf(−xn)→0x_n f(-x_n) \to 0 as n→∞n \to \infty.