Problem 31
Question
Use any test to determine whether the series is absolutely convergent, conditionally convergent, or divergent. \( \displaystyle \sum_{n = 2}^{\infty} \frac {( - 1)^n}{\ln n} \)
Step-by-Step Solution
Verified Answer
The series is conditionally convergent.
1Step 1: Determine Absolute Convergence
To check if the series is absolutely convergent, consider the series of the absolute values:\[ \sum_{n=2}^{\infty} \left| \frac{(-1)^n}{\ln n} \right| = \sum_{n=2}^{\infty} \frac{1}{\ln n}. \]The series \( \sum \frac{1}{\ln n} \) isn't a standard p-series or geometric series, so we need to compare it with a known divergent series or find its behavior. The integral test or limit comparison test can help establish the divergence of this series.
2Step 2: Use Divergence Test for Absolute Series
Since \( \ln n \to \infty \) as \( n \to \infty \), the terms \( \frac{1}{\ln n} \) do not tend to zero fast enough. We can't directly apply the divergence test here, so instead, apply the integral test.Consider the integral:\[ \int_{2}^{\infty} \frac{1}{\ln x} \, dx. \]This integral diverges because the integral of \( \frac{1}{\ln x} \) from 2 to infinity is infinite. Therefore, by the integral test, \( \sum \frac{1}{\ln n} \) diverges. Hence, the original series is not absolutely convergent.
3Step 3: Check Conditional Convergence Using Alternating Series Test
For the alternating series test, consider the original series:\[ \sum_{n=2}^{\infty} \frac{(-1)^n}{\ln n}. \]This is an alternating series with \( b_n = \frac{1}{\ln n}. \)- Check if \( b_n \) is decreasing: Since \( \ln n \) is increasing for \( n \geq 2 \), \( b_n = \frac{1}{\ln n} \) is decreasing.- Check if \( \lim_{n \to \infty} b_n = 0 \): Yes, \( \lim_{n \to \infty} \frac{1}{\ln n} = 0. \)Both conditions of the alternating series test are satisfied, so the series is conditionally convergent.
Key Concepts
Alternating Series TestIntegral TestConditional ConvergenceDivergence Test
Alternating Series Test
The Alternating Series Test is a handy tool when dealing with series where the terms change signs. Let's consider the series \( \sum_{n=2}^{\infty} \frac{(-1)^n}{\ln n} \). An alternating series can be formally described by an expression like \( \sum (-1)^n a_n \), where \( a_n \) is always positive. For the Alternating Series Test to confirm convergence, we need two key conditions:
- The sequence \( a_n \) must be decreasing. Essentially, each term should be smaller than or equal to the one before it. In our series, the term \( a_n = \frac{1}{\ln n} \) decreases as \( \ln n \) grows with \( n \).
- The limit of \( a_n \) as \( n \) approaches infinity should be zero. For our series, \( \lim_{n \to \infty} \frac{1}{\ln n} = 0 \).
Integral Test
The Integral Test helps to determine the convergence or divergence of a series by relating it to an improper integral. Take the series \( \sum_{n=2}^{\infty} \frac{1}{\ln n} \), which is the absolute value version of our original series. By performing an integral test, we look at the function \( f(x) = \frac{1}{\ln x} \), which needs to be positive, continuous, and decreasing for \( x \geq 2 \).
Set up the improper integral \( \int_{2}^{\infty} \frac{1}{\ln x} \, dx \). If this integral converges, then the series converges. Conversely, if the integral diverges, so does the series.
Upon evaluating, it turns out that the integral \( \int_{2}^{\infty} \frac{1}{\ln x} \, dx \) diverges. This means our series of absolute values, \( \sum \frac{1}{\ln n} \), also diverges, indicating that the original series is not absolutely convergent.
Set up the improper integral \( \int_{2}^{\infty} \frac{1}{\ln x} \, dx \). If this integral converges, then the series converges. Conversely, if the integral diverges, so does the series.
Upon evaluating, it turns out that the integral \( \int_{2}^{\infty} \frac{1}{\ln x} \, dx \) diverges. This means our series of absolute values, \( \sum \frac{1}{\ln n} \), also diverges, indicating that the original series is not absolutely convergent.
Conditional Convergence
Conditional Convergence arises when a series converges, but not absolutely. This means that while the original series converges, the series of its absolute values does not. Let's look at our series, \( \sum_{n=2}^{\infty} \frac{(-1)^n}{\ln n} \).
We've already determined through the Alternating Series Test that it is convergent. But when considering the absolute values series, \( \sum \frac{1}{\ln n} \), we found that it diverges. This combination — where the series itself converges, but its absolute counterpart does not — is the hallmark of conditional convergence.
Conditional convergence reveals that the series relies on the interplay of positive and negative terms to "cancel out" and remain finite as \( n \) increases, offering a deeper insight into the behavior of alternating series.
We've already determined through the Alternating Series Test that it is convergent. But when considering the absolute values series, \( \sum \frac{1}{\ln n} \), we found that it diverges. This combination — where the series itself converges, but its absolute counterpart does not — is the hallmark of conditional convergence.
Conditional convergence reveals that the series relies on the interplay of positive and negative terms to "cancel out" and remain finite as \( n \) increases, offering a deeper insight into the behavior of alternating series.
Divergence Test
The Divergence Test is one of the simplest methods for checking a series for divergence. It states that if the limit of the sequence terms \( a_n \) as \( n \) approaches infinity is not zero, the series definitely diverges. However, if that limit equals zero, the test provides no conclusive result because the series may still either converge or diverge.
In our problem, checking the series \( \sum_{n=2}^{\infty} \frac{1}{\ln n} \), the term \( \frac{1}{\ln n} \) does indeed approach zero as \( n \) heads to infinity.
Since the limit is zero, we can’t conclude anything about convergence from the Divergence Test. That’s why other methods like the Integral Test are employed. Thus, while the Divergence Test is often the first step, it's not always the final say in determining a series' behavior.
In our problem, checking the series \( \sum_{n=2}^{\infty} \frac{1}{\ln n} \), the term \( \frac{1}{\ln n} \) does indeed approach zero as \( n \) heads to infinity.
Since the limit is zero, we can’t conclude anything about convergence from the Divergence Test. That’s why other methods like the Integral Test are employed. Thus, while the Divergence Test is often the first step, it's not always the final say in determining a series' behavior.
Other exercises in this chapter
Problem 31
If \( k \) is a positive integer, find the radius of convergence of the series \( \sum_{n = 0}^{\infty} \frac {(n!)^k}{(kn)!} x^n \)
View solution Problem 31
Test the series for convergence or divergence. \( \displaystyle \sum_{k= 1}^{\infty} \frac {5^k}{3^k + 4^k} \)
View solution Problem 31
Is the 50th partial sum \( s_{50} \) of the alternating series \( \displaystyle \sum_{n = 1}^{\infty} (-1)^{n - 1} /n \) an overestimate or an underestimate of
View solution Problem 31
Determine whether the series converges or diverges. \( \displaystyle \sum_{n = 1}^{\infty} \sin \left( \frac {1}{n} \right) \)
View solution