In the last section we introduced mathematical induction by looking at proofs involving summation formulas. In this section we look at some more examples, including divisibility proofs and inequality proofs.
As in the previous section, we will assume \(n\in \mathbb{Z}\text{.}\) We will continue to use the same Structure of a Mathematical Induction Proof. Before looking at new examples of induction proofs, letβs make sure we can write a summation proof.
Our next example looks at using induction to prove a divisibility statement. We proved several properties of divisibility in SectionΒ 4.3. Now we want to be able to use those properties to simplify our induction proofs. Remember, a written proof just needs to make a clear series of logical steps. It does not need to explain how the writer thought to do those steps.
Scratchwork: before proving the induction step, we might need to do some scratchwork. This is not part of the proof, but will help us structure a clear proof. We can do some algebra on the expression we in the βshowβ statement:
\begin{align*}
3&\mid (k^3-k)\ \ \textrm{ by the induction assumption}\\
3&\mid (k^3-k)+3k^2+3k\ \ \textrm{ since we just added multiples of 3}\\
3&\mid (k^3-k)+3k^2+3k+1-1\ \ \textrm{ since 1-1 is 0}\\
3&\mid k^3+3k^2+3k+1-k-1\ \ \textrm{ rearranged terms}\\
3&\mid ((k+1)^3-(k+1)).
\end{align*}
The point of the proof is that the reader can follow all of the steps. You do not need to explain where your choices came from. For example, in the proof in ExampleΒ 5.3.1, it might not be obvious why you should add \(3k^2+3k\text{,}\) but the reader just needs to know that 3 divides the expression, which is clear.
Divisibility proofs in this form can be challenging to do at first, but with practice, they often end up being some of the easier induction proofs to do, as they all follow a similar format.
\begin{align*}
2^{2(k+1)}-1&=2^{2k+2}-1\\
&=(2^2)(2^{2k})-1\ \ \textrm{ used exponent rules to rewrite}\\
&=(4)(2^{2k})-1\\
&=(3+1)(2^{2k})-1\ \ \textrm{ maybe not an obvious step,}\\
&\ \ \ \ \ \ \ \textrm{ but we are looking for multiples of 3}\\
&=3\cdot 2^{2k}+2^{2k}-1\ \ \textrm{ distribute}
\end{align*}
But by our assumption, \(3\mid 2^{2k}-1\) and since the remaining term is a multiple of 3, \(3\mid 3\cdot 2^{2k}+2^{2k}-1\text{.}\)
\begin{align*}
3&\mid 2^{2k}-1\ \ \textrm{ by the induction assumption}\\
3&\mid 2^{2k}-1+3\cdot 2^{2k}\ \ \textrm{ since we just added a multiple of 3}\\
3&\mid (3+1)(2^{2k})-1\ \ \textrm{ factored out } 2^{2k}\\
3&\mid (4)(2^{2k})-1\\
3&\mid (2^2)(2^{2k})-1\\
3&\mid (2^{2k+2})-1\\
3&\mid (2^{2(k+1)})-1
\end{align*}
Now we apply induction to statements involving inequalities. These will be some of the most challenging induction proofs as they often take a bit more trial and error in the scratchwork. When dealing with equalities, you know that you have to add the same thing to both sides. However, with inequalities you can change one side as long as you maintain the inequality. For example, if you know \(P>Q\text{,}\) then you also know \(P+2>Q\text{.}\) The fact that we can change one side of the expression makes it more challenging to see what steps to take.
Now we want to make a useful observation: in the LHS expression we are addding. In the RHS expression we are multiplying. We need a way to go between them. One way is to notice that \(2(A)=A+A\text{.}\) Then
Evaluate the sum \(\sum_{i=1}^{n}\frac{i}{(i+1)!}\) for all \(n=1, 2, 3, 4, 5\text{.}\) Make a conjecture for a formula for the sum for a general \(n\text{,}\) and prove your conjecture by induction.
You can only do induction on integers, and only one integer at a time. So in this problem you want to do induction on \(n\text{,}\) not \(x\text{.}\) Let \(x\) just be a variable. This problem does require \(x>-1\text{,}\) so make sure you indicate in your proof where you needed to use that condition.