2

I'm starting to understand how induction works (with the whole $k \to k+1$ thing), but I'm not exactly sure how summations play a role. I'm a bit confused by this question specifically:

$$ \sum_{i=1}^n 3i-2 = \frac{n(3n-1)}{2} $$

Any hints would be greatly appreciate

4 Answers4

7

Base case:

  • Let $n=1$ and test: $$\sum_{i=1}^1(3i-2)=3-2=1=\frac{1(3\cdot 1-1)}{2}$$

  • True for $n = 1$

Induction Hypothesis:

  • Assume that it is true for $k$: assume that $$\sum_{i=1}^k(3i-2)=\frac{k(3k-1)}{2}.$$

Inductive Step:

  • Prove, using the Inductive Hypothesis as a premise, that $$\sum_{i=1}^{k+1}(3i-2)=\left(\sum_{i = 1}^k(3i - 2)\right) + (3(k+1)-2) = \frac{(k+1)(3(k+1)-1)}{2}.$$
amWhy
  • 209,954
  • So do I just manipulate the right side of the equation to try to get it to look like the original k(3k-1)/2 ? – K. Barresi Jan 25 '13 at 01:58
  • No, manipulate the inner third (in the equality chain of last line) to get the right hand side. You know, from the inductive hypothesis, what that the sum $$\left(\sum_{i = 1}^k(3i - 2)\right) = \frac{k(3k - 1)}{2}.$$ Add that to the term $3(k+1)- 2$ to obtain the right hand side. The right hand side is precisely the form you want to exhibit to show that the given sum holds for $k+1$ if it holds for $k$, hence, you will have proven, by induction, that for all $n$, $$ \sum_{i=1}^n 3i-2 = \frac{n(3n-1)}{2} $$ – amWhy Jan 25 '13 at 02:01
  • Ahh ok I think I understand now. Thank you for your help. – K. Barresi Jan 25 '13 at 02:19
  • Sorry to bug you again, but I appear to be stuck. I've reached:

    $$ (3k-2)+(3[k+1]-2)=\frac{k+1(3[k+1]-1)}{2} $$ which I simplified to $$ 12k-2=3k^2+5k+2 $$

    – K. Barresi Jan 25 '13 at 04:05
  • That's exactly what you want: you want (k+1) in every spot where k occurs in the inductive hypothesis: i.e. you want the right hand side of my last line. That shows that what holds for k, holds when every k is replaced by (k+1): you want the same "form" as on the right hand side. – amWhy Jan 25 '13 at 04:12
  • Oh so $12k-2=3k^2+5k+2$ is actually showing that? I must be a bit rusty with my algebra ;) – K. Barresi Jan 25 '13 at 04:19
  • The first expression before simplified, is all you need to show. No need to simplify. – amWhy Jan 25 '13 at 04:30
  • Oh okay. Thanks for all your help amWhy – K. Barresi Jan 25 '13 at 04:33
2

If we mimic amWhy's proof more generally we obtain a powerful result on telescoping sums:

$$ \sum_{i=1}^n f(i) =?\ g(n)$$

Base case:

  • Let $n=1$ and test: $$\sum_{i=1}^1 f(i) = f(1) \color{#c00}{=?}\ g(1)$$

  • The $ $ Base Case $\ \,n = 1\ $ holds true $\iff \color{#C00}{f(1) = g(1)}$

Induction Hypothesis:

  • Assume that it is true for $\, n = k$: assume that $$\sum_{i=1}^k f(i) = g(k).$$

Inductive Step:

  • Prove, using the Induction Hypothesis as a premise, that $$\sum_{i=1}^{k+1}f(i)=\left(\sum_{i=1}^k f(i)\right) + f(k\!+\!1) = g(k) + f(k\!+\!1) \color{#0a0}{=?}\ g(k\!+\!1).$$

  • The Inductive Step from $k$ to $k+1$ is true $\iff \color{#0a0}{ g(k\!+\!1) - g(k) = f(k\!+\!1)}$

Therefore we have proved by induction the following generic summation criterion

$\displaystyle\quad\sum_{i=1}^n f(i) = g(n)\iff \color{#c00}{g(1) = f(1)}\,\ {\rm and}\,\ \color{#0a0}{g(n\!+\!1)-g(n) = f(n\!+\!1)}\ $ for $\,n \ge 1$

This theorem reduces the inductive proof to simply verifying the $\rm\color{#c00}{RHS}$ $\rm\color{#0a0}{equalities}$, which is trivial polynomial arithmetic when $f(n),g(n)$ are polynomials in $n,\,$ so trivial it can be done purely mechanically by a high-school student (or computer). No insight (magic) is required - no rabbits need be pulled from a hat.

The above theorem is an example of telescopy, also known as the Fundamental Theorem of Difference Calculus, depending on context. You can find many more examples of telescopy and related results in other answers here.

Bill Dubuque
  • 272,048
Math Gems
  • 19,574
2

I will prove it in two way for you:

1- Mathematical Induction:

If $n=1$ then the left side is $1$ and also the right side is $1$ too.

Now think that we have $\sum_{i=1}^{n}(3i-2)=\frac{n(3n-1)}{2}$, we should show $\sum_{i=1}^{n+1}(3i-2)=\frac{(n+1)(3(n+1)-1)}{2}$ that is $\sum_{i=1}^{n+1}(3i-2)=\frac{(n+1)(3n+2)}{2}$. But we can write; $$\sum_{i=1}^{n+1}(3i-2)=(\sum_{i=1}^{n}(3i-2))+(3(n+1)-2)=\frac{n(3n-1)}{2}+(3n+1)=\frac{3n^{2}-n+6n+2}{2}=\frac{3n^{2}+5n+2}{2}=\frac{3n^{2}+3n+2n+2}{2}=\frac{3n(n+1)+2(n+1)}{2}=\frac{(n+1)(3n+2)}{2}$$. And it finished our work.

2- Without Mathematical Induction:

We know $\sum_{i=1}^{k}i=\frac{k(k+1)}{2}$ and $\sum_{i=1}^{k}c=kc$ for a constant number "c".

Now $$\sum_{i=1}^{n}(3i-2)=3\sum_{i=1}^{n}i-\sum_{i=1}^{n}2=3\frac{n(n+1)}{2}-2n=\frac{n(3n+3-4)}{2}=\frac{n(3n-1)}{2}$$.

omid saba
  • 144
1

For $n=1$ you have $\sum_{i=1}^1(3i-2)=3-2=1=\frac{1(3\cdot 1-1)}{2}$.

Suppose that $$\sum_{i=1}^n(3i-2)=\frac{n(3n-1)}{2}$$ and prove that $$\sum_{i=1}^{n+1}(3i-2)=\frac{(n+1)(3(n+1)-1)}{2}.$$

Sigur
  • 6,416
  • 3
  • 25
  • 45