Hurray 2016!

In the expansion of ( 3 x + 2 ) 2016 (3x+2)^{2016} , what is the exponent of x x in the term with the largest coefficient?


The answer is 1210.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

4 solutions

Let T r T_{r} denote the co-efficient of r t h r^{th} term in the expansion of ( 3 x + 2 ) 2016 \left(3x+2\right)^{2016}
T r + 1 = ( 2016 r ) 3 2016 r 2 r T_{r+1} = \dbinom{2016}{r}3^{2016-r}2^{r}

Replacing r by r-1,
T r = ( 2017 r r 1 ) 3 2017 r 2 r 1 T_{r} = \dbinom{2017-r}{r-1}3^{2017-r}2^{r-1}
T r + 1 T r = 2 ( 2017 r ) 3 r . . . ( 1 ) \dfrac{T_{r+1}}{T_{r}} = \dfrac{2(2017-r)}{3r} ...(1)

Solving for T r + 1 T r 1 \dfrac{T_{r+1}}{T_{r}} \ge 1 yields r 4034 5 r \le \dfrac{4034}{5}
r 806.8 \therefore r \le 806.8

Replacing r by r+1 in ( 1 ) (1) ,

T r + 2 T r + 1 = 2 ( 2016 r ) 3 ( r + 1 ) \dfrac{T_{r+2}}{T_{r+1}} = \dfrac{2(2016-r)}{3(r+1)}

Solving for T r + 2 T r + 1 1 \dfrac{T_{r+2}}{T_{r+1}} \le 1 yields r 4029 5 r \ge \dfrac{4029}{5}
r 805.8 \therefore r \ge 805.8
The only integer satisfying both conditions is 806.
r = 806 \therefore r = 806

Hence the greatest co-efficient term is ( r + 1 ) = 806 + 1 = 807 (r+1) = 806+1 = 807 term.
Co-efficient of x in 80 7 t h 807^{th} term = 2016 806 = 1210 = 2016-806 = 1210

Wow,thank you!

Rohit Udaiwal - 5 years, 5 months ago

Log in to reply

@rohit udaiwal I hope that was well explained.

A Former Brilliant Member - 5 years, 5 months ago

General problem : For what power x k x^k does the expansion of ( a x + b ) n (ax + b)^n have the largest coefficient?

Solution : The coefficient of x k x^k is C k = ( n k ) a k b n k = n ( n k + 1 ) k 1 a k b n k . C_k = \left(\begin{array}{c} n \\ k\end{array}\right) a^k b^{n-k} = \frac{n\cdots(n-k+1)}{k\cdots 1} a^k b^{n-k}. This coefficient is greater than the previous one iff C k > C k 1 n ( n k + 1 ) k 1 a k b n k > n ( n k + 2 ) ( k 1 ) 1 a k 1 b n k + 1 n k + 1 k > b a n + 1 k > b + a a k < ( n + 1 ) a b + a . C_k > C_{k-1} \\ \frac{n\cdots(n-k+1)}{k\cdots 1} a^k b^{n-k} > \frac{n\cdots(n-k+2)}{(k-1)\cdots 1} a^{k-1} b^{n-k+1} \\ \frac{n-k+1}{k} > \frac b a \\ \frac{n+1}{k} > \frac{b+a}a \\ k < (n+1) \frac a{b+a}. The greatest coefficient corresponds to the greatest integer k k for which this is the case, so k = ( n + 1 ) a b + a . k = \left\lfloor (n+1) \frac a{b+a}\right\rfloor.

In this case: k = 2017 3 5 = 1210 . k = \left\lfloor 2017\cdot \frac 3 5 \right\rfloor = \boxed{1210}.

Interesting fact: If we divide the entire expression by 5 2016 5^{2016} we get ( 3 5 x + 2 5 ) 2016 (\tfrac35x + \tfrac25)^{2016} , in which the coefficient of x k x_k corresponds to the probability of having k k successes in an 2016 2016 -times repeated experiment with success probability p = 3 5 p = \tfrac35 . The "peak" of this distribution, which corresponds to the greatest exponent, is the expectation value of this distribution, E = n p = 2016 3 5 1210. \mathbb E = np = 2016\cdot \tfrac35 \approx 1210. This is how I originally guessed the answer to this question. Of course, the derivation above is more "solid" than this intuitive approach.

Nice way to bring in the Bernoulli trials.

I have a doubt, shouldn't it be,
k ( n + 1 ) a b + a k \leq \ \dfrac{(n+1)a}{b+a} ??

A Former Brilliant Member - 5 years, 5 months ago

Log in to reply

If the equal sign applies, then coefficients C k C_k and C k 1 C_{k-1} are equal.

Arjen Vreugdenhil - 5 years, 5 months ago

Log in to reply

Yes. I only pointed that out, because in the next step you mentioned the floor function, and equality would hold if ( b + a ) a ( n + 1 ) (b+a) | a(n+1)

A Former Brilliant Member - 5 years, 5 months ago

Log in to reply

@A Former Brilliant Member I had realized that but didn't want to clog up by solution. :)

Arjen Vreugdenhil - 5 years, 5 months ago
Christopher Boo
Jan 2, 2016

There are nice solutions provided by both @Vighnesh Shenoy and @Manuel Kahayon. I only wanted to show the generalization to this problem by treating everything symbolically :

What is the exponent of x x that have the greatest coefficient in the expansion of ( a x + b ) N (ax+b)^{N} ?

The approach is not far from how you would solve the previous problem, you will arrive at

a ( N + 1 ) a + b \displaystyle \bigg \lfloor \frac{a(N+1)}{a+b}\bigg \rfloor

What we need to be aware of is the sign of a a or b b . If exactly one of them is negative, we have to check if their corresponding index is even. If not, we have to check which neighboring coefficient is larger. In the case where we introduced negatives, it is necessary to change the expression to :

a ( N + 1 ) a + b \displaystyle \bigg \lfloor \frac{|a|(N+1)}{|a|+|b|}\bigg \rfloor

Did you mean, "What is the exponent of x in greatest term of the expansion ( a x + b ) N \left(ax + b\right)^{N} ?" Nice generalisation.

A Former Brilliant Member - 5 years, 5 months ago

Log in to reply

Oops, thanks for the notice.

Christopher Boo - 5 years, 5 months ago

Similar to what @Akhil Bansal wrote with a tweak to it.

A Former Brilliant Member - 5 years, 5 months ago
Manuel Kahayon
Dec 31, 2015

Since the first term is ( 3 x ) 2016 (3x)^{2016} , and the nth term is equal to ( n 1 t h t e r m ) ( 2018 n ) ( 2 ) ( 3 ) ( n 1 ) \frac{(n-1th term)(2018-n)(2)}{(3)(n-1)} , then as long as ( 2018 n ) ( 2 ) ( 3 ) ( n 1 ) 1 \frac{(2018-n)(2)}{(3)(n-1)} \geq 1 , then the nth term will be greater than the n-1th term. Equating this, we get

( 2018 n ) ( 2 ) ( 3 ) ( n 1 ) 1 \frac{(2018-n)(2)}{(3)(n-1)} \geq 1

807.8 n 807.8 \geq n

Rounding off, the 807th term has the largest coefficient, meaning, the 807th term has 1210 1210 is the exponent of x in the 807th term.

We can directly use the formula,
Numerically greatest term in the expansion of ( 1 + x ) n (1+x)^n is ( n + 1 ) x 1 + x = m \dfrac{(n+1)|x|}{1+|x|} = m where m \lfloor m \rfloor is exponent of x

Akhil Bansal - 5 years, 5 months ago

Log in to reply

Oh, thanks, didn't know that...

Sorta disappointed to know that there was already a formula for this...

Manuel Kahayon - 5 years, 5 months ago

Log in to reply

@Akhil Bansal The formula you provided gives the numerically greatest term, that is, along with the input for value of x. But, the question asked the term with greatest co-efficient which may or may not be the same.

A Former Brilliant Member - 5 years, 5 months ago

Log in to reply

@A Former Brilliant Member Yes, formula also tells about which term contains numerically greatest term.That's how I finded exponent of x but I made one assumption that x = 1.

Akhil Bansal - 5 years, 5 months ago

Why is it not 2016/2=1008,and can you provide a proof of this formula?

Rohit Udaiwal - 5 years, 5 months ago

Log in to reply

Well, I do know that if it is the expansion of ( n x + n ) 2016 (nx+n)^{2016} for some n then the 1008th term would be the greatest ( x C x 2 x C \frac{x}{2} is the greatest for some x), but since there are different coefficients (i.e. ( a x + b ) 2016 (ax+b)^{2016} , the terms with the largest coefficient tends to "shift" towards the left or the right, either left if a is greater than b or vice versa.

(Sorry, i kinda suck at explaining)

(Also, why is this in combinatorics not algebra?)

Manuel Kahayon - 5 years, 5 months ago

Log in to reply

@Manuel Kahayon It is in combinatorics because it involved the binomial co-efficients.

A Former Brilliant Member - 5 years, 5 months ago

it doesn't satisfy the inequalities: 2n+2>5k and 2n-3<5k . 807 is the only value that do

Benjamin ononogbu - 5 years, 5 months ago

Sorry, I didn't know the proof. I studies it in 2014.

Akhil Bansal - 5 years, 5 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...