Difference between revisions of "2015 AMC 10B Problems/Problem 23"
(→Solution 4) |
m (→Problem) |
||
(49 intermediate revisions by 11 users not shown) | |||
Line 1: | Line 1: | ||
==Problem== | ==Problem== | ||
− | Let <math>n</math> be a positive integer greater than 4 such that the decimal representation of <math>n!</math> ends in <math>k</math> zeros and the decimal representation of <math>(2n)!</math> ends in <math>3k</math> zeros. Let <math>s</math> denote the sum of the four least possible values of <math>n</math>. What is the sum of the digits of <math>s</math>? | + | Let <math>n</math> be a positive integer greater than <math>4</math> such that the decimal representation of <math>n!</math> ends in <math>k</math> zeros and the decimal representation of <math>(2n)!</math> ends in <math>3k</math> zeros. Let <math>s</math> denote the sum of the four least possible values of <math>n</math>. What is the sum of the digits of <math>s</math>? |
<math> \textbf{(A) }7\qquad\textbf{(B) }8\qquad\textbf{(C) }9\qquad\textbf{(D) }10\qquad\textbf{(E) }11 </math> | <math> \textbf{(A) }7\qquad\textbf{(B) }8\qquad\textbf{(C) }9\qquad\textbf{(D) }10\qquad\textbf{(E) }11 </math> | ||
− | + | ==Solution 1== | |
− | |||
− | |||
A trailing zero requires a factor of two and a factor of five. Since factors of two occur more often than factors of five, we can focus on the factors of five. We make a chart of how many trailing zeros factorials have: | A trailing zero requires a factor of two and a factor of five. Since factors of two occur more often than factors of five, we can focus on the factors of five. We make a chart of how many trailing zeros factorials have: | ||
Line 18: | Line 16: | ||
Secondly, we look at the case when <math>n!</math> has <math>2</math> zeros and <math>(2n)!</math> has <math>6</math> zeros. If <math>n=10,11,12</math>, <math>(2n)!</math> has only <math>4</math> zeros. But for <math>n=13,14</math>, <math>(2n)!</math> has <math>6</math> zeros. Thus, the smallest four values of <math>n</math> that work are <math>n=8,9,13,14</math>, which sum to <math>44</math>. The sum of the digits of <math>44</math> is <math>\boxed{\mathbf{(B)\ }8}</math> | Secondly, we look at the case when <math>n!</math> has <math>2</math> zeros and <math>(2n)!</math> has <math>6</math> zeros. If <math>n=10,11,12</math>, <math>(2n)!</math> has only <math>4</math> zeros. But for <math>n=13,14</math>, <math>(2n)!</math> has <math>6</math> zeros. Thus, the smallest four values of <math>n</math> that work are <math>n=8,9,13,14</math>, which sum to <math>44</math>. The sum of the digits of <math>44</math> is <math>\boxed{\mathbf{(B)\ }8}</math> | ||
− | + | ==Solution 2== | |
− | By Legendre's Formula and the information given, we have that <math>3\left(\left\lfloor{\frac{n}{5}}\right\rfloor+\left\lfloor{\frac{n}{25}}\right\rfloor\right)=\left\lfloor{\frac{2n}{5}}\right\rfloor+\left\lfloor{\frac{2n}{25}}\right\rfloor</math>. | + | By [[Legendre's Formula]] and the information given, we have that <math>3\left(\left\lfloor{\frac{n}{5}}\right\rfloor+\left\lfloor{\frac{n}{25}}\right\rfloor\right)=\left\lfloor{\frac{2n}{5}}\right\rfloor+\left\lfloor{\frac{2n}{25}}\right\rfloor</math>. |
− | + | We have <math>n<100</math> as there is no way that if <math>n>100</math>, <math>(2n)!</math> would have <math>3</math> times as many zeroes as <math>n!</math>. | |
− | First, let's plug in the number <math>5</math> | + | First, let's plug in the number <math>5</math>. |
We get that <math>3(1)=1</math>, which is obviously not true. Hence, <math>n>5</math> | We get that <math>3(1)=1</math>, which is obviously not true. Hence, <math>n>5</math> | ||
After several attempts, we realize that the RHS needs <math>1</math> to <math>2</math> more "extra" zeroes than the LHS. Hence, <math>n</math> is greater than a multiple of <math>5</math>. | After several attempts, we realize that the RHS needs <math>1</math> to <math>2</math> more "extra" zeroes than the LHS. Hence, <math>n</math> is greater than a multiple of <math>5</math>. | ||
− | + | We find that the least four possible <math>n</math> are <math>8,9,13,14</math>. | |
<math>8+9+13+14=17+27=44\implies 4+4=8\implies\boxed{B}</math>. | <math>8+9+13+14=17+27=44\implies 4+4=8\implies\boxed{B}</math>. | ||
− | |||
− | |||
+ | == Solution 3 == | ||
+ | |||
+ | Let <math>n=5m+k</math> for some natural numbers <math>m</math>, <math>k</math> such that <math>k\in\{0,1,2,3,4\}</math>. Notice that <math>n<5^3=125</math>. Thus | ||
+ | <cmath>3(\left\lfloor\frac{n}{5}\right\rfloor+\left\lfloor\frac{n}{25}\right\rfloor)=\left\lfloor\frac{2n}{5}\right\rfloor+\left\lfloor\frac{2n}{25}\right\rfloor+\left\lfloor\frac{2n}{125}\right\rfloor</cmath> | ||
+ | For smaller <math>n</math>, we temporarily let <math>\left\lfloor\frac{2n}{125}\right\rfloor=0</math> | ||
+ | <cmath>3(\left\lfloor\frac{n}{5}\right\rfloor+\left\lfloor\frac{n}{25}\right\rfloor)=\left\lfloor\frac{2n}{5}\right\rfloor+\left\lfloor\frac{2n}{25}\right\rfloor</cmath> | ||
+ | <cmath>3(\left\lfloor\frac{5m+k}{5}\right\rfloor+\left\lfloor\frac{5m+k}{25}\right\rfloor)=\left\lfloor\frac{2(5m+k)}{5}\right\rfloor+\left\lfloor\frac{2(5m+k)}{25}\right\rfloor</cmath> | ||
+ | <cmath>3(\left\lfloor\frac{5m+k}{5}\right\rfloor+\left\lfloor\frac{5m+k}{25}\right\rfloor)=\left\lfloor\frac{10m+2k}{5}\right\rfloor+\left\lfloor\frac{10m+2k}{25}\right\rfloor</cmath> | ||
+ | <cmath>3m+3\left\lfloor\frac{5m+k}{25}\right\rfloor=2m+\left\lfloor\frac{2k}{5}\right\rfloor+\left\lfloor\frac{10m+2k}{25}\right\rfloor</cmath> | ||
+ | <cmath>m+3\left\lfloor\frac{5m+k}{25}\right\rfloor=\left\lfloor\frac{2k}{5}\right\rfloor+\left\lfloor\frac{10m+2k}{25}\right\rfloor</cmath> | ||
+ | To minimize <math>n</math>, we let <math>\left\lfloor\frac{5m+k}{25}\right\rfloor=\left\lfloor\frac{10m+2k}{25}\right\rfloor=0</math>, then | ||
+ | <cmath>m=\left\lfloor\frac{2k}{5}\right\rfloor</cmath> | ||
+ | Since <math>k<5</math>, <math>m>0</math>, the only integral value of <math>m</math> is <math>1</math>, from which we have <math>k=3,4\Longrightarrow n=8,9</math>. | ||
+ | |||
+ | Now we let <math>\left\lfloor\frac{5m+k}{25}\right\rfloor=0</math> and <math>\left\lfloor\frac{10m+2k}{25}\right\rfloor=1</math>, then | ||
+ | <cmath>m=\left\lfloor\frac{2k}{5}\right\rfloor+\left\lfloor\frac{10m+2k}{25}\right\rfloor</cmath> | ||
+ | Since <math>k<5</math>, <math>10m>15\Longrightarrow m\ge2</math>. | ||
+ | |||
+ | If <math>m>2</math>, then | ||
+ | <cmath>m>\left\lfloor\frac{2k}{5}\right\rfloor+\left\lfloor\frac{10m+2k}{25}\right\rfloor</cmath> | ||
+ | which is a contradiction. | ||
+ | |||
+ | Thus <math>m=2\Longrightarrow\left\lfloor\frac{2k}{5}\right\rfloor=1\Longrightarrow n=13,14</math> | ||
− | == | + | Finally, the sum of the four smallest possible <math>n=8+9+13+14=44</math> and <math>4+4=8</math>. <math>\boxed{\mathrm{(B)}}</math> |
− | + | ~ Nafer | |
− | < | + | |
− | + | == Solution 4== | |
− | < | + | |
− | < | + | We first note that the number of 0's in <math>n!</math> is determined by how many 5's are in the prime factorization. We use Legendre's Formula and split up into two cases: |
+ | |||
+ | <math>\textbf{CASE ONE: } 5\leq 2n < 25.</math> | ||
+ | |||
+ | The only way we can fulfill the requirements is if <math>\lfloor{\dfrac{n}{5}}\rfloor = 1</math> and <math>\lfloor{\dfrac{2n}{5}}\rfloor=3</math> which means that <math>5\leq n <10</math> and <math>15\leq 2n < 20</math>. The only way this works is if <math>n = 8 \text{ or } 9.</math> | ||
+ | |||
+ | <math>\textbf{CASE TWO: } 25 \leq 2n</math>. | ||
+ | |||
+ | Since we want the smallest values of <math>n</math>, we first try it when <math>2n<30.</math> Thus <math>(2n)!</math> has 6 zeros, which implies that <math>n!</math> must have 2. The only way to do this while maintaining our restrictions for <math>2n</math> is if <math>n = 13 \text{ or } 14.</math> | ||
+ | |||
+ | So the sum of the four values is <math>8+9+13+14=44</math> so the digit is sum is <math>\boxed{\mathbf{(B)\ }8}.</math> | ||
+ | |||
+ | -ConfidentKoala4 | ||
+ | |||
+ | |||
+ | == Solution 5 == | ||
+ | |||
+ | We will use trial and error to determine the answer to this problem. If <math>n = 5</math>, then <math>n!</math> has <math>1</math> zero, and <math>2n!</math> will have <math>2</math> zeros. But <math>3 \cdot 1 \neq 2</math> so <math>n = 5</math> does not work. Similarly <math>n = 6, 7</math> do not work either. But <math>n = 8</math> works because <math>8!</math> has <math>1</math> zero and <math>16!</math> has <math>3</math> zeros. Note that <math>n = 9</math> also works because <math>9!</math> has <math>1</math> zero and <math>18!</math> has <math>3</math> zeros. After performing trial and error several times, we find that <math>n = 10, 11, 12</math> do not work but <math>n = 13, 14</math> do work. Therefore, the four smallest values of <math>n</math> are <math>8, 9, 13, 14</math>. Therefore adding them together gives <math>44</math> and our answer is <math>\boxed{\mathbf{(B)\ }8}.</math> | ||
+ | |||
+ | -[[User: Yiyj1|Yiyj1]] | ||
==See Also== | ==See Also== | ||
{{AMC10 box|year=2015|ab=B|num-b=22|num-a=24}} | {{AMC10 box|year=2015|ab=B|num-b=22|num-a=24}} | ||
{{MAA Notice}} | {{MAA Notice}} |
Latest revision as of 00:16, 19 July 2024
Problem
Let be a positive integer greater than such that the decimal representation of ends in zeros and the decimal representation of ends in zeros. Let denote the sum of the four least possible values of . What is the sum of the digits of ?
Solution 1
A trailing zero requires a factor of two and a factor of five. Since factors of two occur more often than factors of five, we can focus on the factors of five. We make a chart of how many trailing zeros factorials have:
We first look at the case when has zero and has zeros. If , has only zeros. But for , has zeros. Thus, and work.
Secondly, we look at the case when has zeros and has zeros. If , has only zeros. But for , has zeros. Thus, the smallest four values of that work are , which sum to . The sum of the digits of is
Solution 2
By Legendre's Formula and the information given, we have that .
We have as there is no way that if , would have times as many zeroes as .
First, let's plug in the number . We get that , which is obviously not true. Hence,
After several attempts, we realize that the RHS needs to more "extra" zeroes than the LHS. Hence, is greater than a multiple of .
We find that the least four possible are .
.
Solution 3
Let for some natural numbers , such that . Notice that . Thus For smaller , we temporarily let To minimize , we let , then Since , , the only integral value of is , from which we have .
Now we let and , then Since , .
If , then which is a contradiction.
Thus
Finally, the sum of the four smallest possible and .
~ Nafer
Solution 4
We first note that the number of 0's in is determined by how many 5's are in the prime factorization. We use Legendre's Formula and split up into two cases:
The only way we can fulfill the requirements is if and which means that and . The only way this works is if
.
Since we want the smallest values of , we first try it when Thus has 6 zeros, which implies that must have 2. The only way to do this while maintaining our restrictions for is if
So the sum of the four values is so the digit is sum is
-ConfidentKoala4
Solution 5
We will use trial and error to determine the answer to this problem. If , then has zero, and will have zeros. But so does not work. Similarly do not work either. But works because has zero and has zeros. Note that also works because has zero and has zeros. After performing trial and error several times, we find that do not work but do work. Therefore, the four smallest values of are . Therefore adding them together gives and our answer is
See Also
2015 AMC 10B (Problems • Answer Key • Resources) | ||
Preceded by Problem 22 |
Followed by Problem 24 | |
1 • 2 • 3 • 4 • 5 • 6 • 7 • 8 • 9 • 10 • 11 • 12 • 13 • 14 • 15 • 16 • 17 • 18 • 19 • 20 • 21 • 22 • 23 • 24 • 25 | ||
All AMC 10 Problems and Solutions |
The problems on this page are copyrighted by the Mathematical Association of America's American Mathematics Competitions.