0
$\begingroup$

Question : Let $a_r$ denote the following $$(1+x+x^2)^n=\sum_{r=0}^{2n}a_rx^r$$ then prove the following

$$\sum_{r=0}^{n}(-1)^r\binom n r a_r = \begin{cases} 0 & n \ne 3k \text{ for all } k\in\mathbb{N} \\ (-1)^\frac{4n}{3} \dbinom n {\frac{n}{3}} & n = 3k \text{ for all } k\in\mathbb{N} \end{cases}$$

My Attempt:

$$\sum_{r=0}^{n}(-1)^r\binom n r (1+x+x^2)^n$$ where coeffiecient of $x^r$ is the answer

$$(1+x+x^2)^n\sum_{r=0}^{n}(-1)^r\binom n r=0$$ hence my answer comes out to be $0$ regardless of whether $n$ is divisible by 3 or not

another similar approach I could think of is

$$(1-x+x^2)=\sum_{r=0}^{2n}(-1)^r a_r x^r$$ hence using this we get the original problem to convert into $$(1-x+x^2)^n \sum_{r=0}^{n}\binom n r= (1-x+x^2)^n \cdot 2^n$$ where coefficient of again $x^r$ is the answer which comes out to be $(-1)^ra_r\cdot2^n $ this time. Where am I going wrong?

$\endgroup$
2
  • $\begingroup$ I do not understand "where coeffiecient of $x^r$ is the answer". The coefficient of $x^r$ in $\sum_{r=0}^{n}(-1)^r\binom n r (1+x+x^2)^n=\sum_{k=0}^{n}(-1)^k\binom nk (1+x+x^2)^n$ is $\sum_{k=0}^{n}(-1)^k\binom nk a_r=0a_r=0$ but this gives no information about $a_r.$ $\endgroup$ Commented Oct 1, 2023 at 10:00
  • 1
    $\begingroup$ Spoiler. $\endgroup$ Commented Oct 1, 2023 at 10:06

0

You must log in to answer this question.