Difference between revisions of "1998 AIME Problems/Problem 14"
m (→Solution: failed to parse error) |
(→Solution) |
||
Line 3: | Line 3: | ||
== Solution == | == Solution == | ||
− | <div style="text-align:center;"><math> | + | <div style="text-align:center;"><math>2mnp = (m+2)(n+2)(p+2)</math></div> |
Let’s solve for <math>p</math>: | Let’s solve for <math>p</math>: | ||
Line 15: | Line 15: | ||
<div style="text-align:center;"><math>p = \frac{2(m+2)(n+2)}{(m-2)(n-2) - 8}</math></div> | <div style="text-align:center;"><math>p = \frac{2(m+2)(n+2)}{(m-2)(n-2) - 8}</math></div> | ||
− | Clearly, we want to minimize the denominator, so <math> | + | Clearly, we want to minimize the denominator, so <math>(m-2)(n-2) - 8 = 1 \Longrightarrow (m-2)(n-2) = 9</math>. The possible pairs of factors of <math>9</math> are <math>(1,9)(3,3)</math>. These give <math>m = 3, n = 11</math> and <math>m = 5, n = 5</math> respectively. Substituting into the numerator, we see that the first pair gives <math>130</math>, while the second pair gives <math>98</math>. We can quickly test for the denominator assuming other values to convince ourselves that <math>1</math> is the best possible value for the denominator. Hence, the solution is <math>130</math>. |
+ | |||
+ | |||
+ | Proof that setting the denominator <math>(m - 2)(n - 2) - 8</math> to <math>1</math> is optimal: Suppose <math>(m - 2)(n - 2) = 9</math>, and suppose for the sake of contradiction that there exist <math>m', n'</math> such that <math>(m' - 2)(n' - 2) = 8 + d</math> for some <math>d > 1</math> and such that | ||
+ | <cmath>\frac{2(m + 2)(n + 2)}{(m - 2)(n - 2) - 8} < \frac{2(m' + 2)(n' + 2)}{(m' - 2)(n' - 2) - 8}.</cmath> | ||
+ | This implies that | ||
+ | <cmath>d(m + 2)(n + 2) < (m' + 2)(n' + 2),</cmath> | ||
+ | and | ||
+ | <cmath>d((m - 2)(n - 2) + 4(m + n)) < (m' - 2)(n' - 2) + 4(m' + n').</cmath> | ||
+ | Substituting gives | ||
+ | <cmath>d(9 + 4(m + n)) < 8 + d + 4(m' + n'),</cmath> | ||
+ | which we rewrite as | ||
+ | <cmath>d(8 + 4(m + n)) < 24 + 4((m' - 2) + (n' - 2)).</cmath> | ||
+ | Next, note that for <math>p'</math> to be positive, we must have <math>m' - 2</math> and <math>n' - 2</math> be positive, so | ||
+ | <cmath>(m' - 2) + (n' - 2) \leq (m' - 2)(n' - 2) + 1 = 9 + d.</cmath> | ||
+ | So | ||
+ | <cmath>d(8 + 4(m + n)) < 8 + 4(9 + d)</cmath> | ||
+ | <cmath>d(4 + 4(m + n)) < 44</cmath> | ||
+ | <cmath>d(1 + m + n) < 11</cmath> | ||
+ | Next, we must have that <math>m - 2</math> and <math>n - 2</math> are positive, so <math>3 \leq m</math> and <math>3 \leq n</math>. Also, <math>2 \leq d</math> by how we defined <math>d</math>. So | ||
+ | <cmath>2(1 + 3 + 3) < 11,</cmath> | ||
+ | a contradiction. We already showed above that there are some values of <math>m</math> and <math>n</math> such that <math>(m - 2)(n - 2) = 9</math> that work, so this proves that one of these pairs of values of <math>m</math> and <math>n</math> must yield the maximal value of <math>p</math>. | ||
== See also == | == See also == |
Revision as of 00:07, 17 September 2012
Problem
An rectangular box has half the volume of an rectangular box, where and are integers, and What is the largest possible value of ?
Solution
Let’s solve for :
For the denominator, we will use a factoring trick (colloquially known as SFFT), which states that .
Clearly, we want to minimize the denominator, so . The possible pairs of factors of are . These give and respectively. Substituting into the numerator, we see that the first pair gives , while the second pair gives . We can quickly test for the denominator assuming other values to convince ourselves that is the best possible value for the denominator. Hence, the solution is .
Proof that setting the denominator to is optimal: Suppose , and suppose for the sake of contradiction that there exist such that for some and such that
This implies that
and
Substituting gives
which we rewrite as
Next, note that for to be positive, we must have and be positive, so
So
Next, we must have that and are positive, so and . Also, by how we defined . So
a contradiction. We already showed above that there are some values of and such that that work, so this proves that one of these pairs of values of and must yield the maximal value of .
See also
1998 AIME (Problems • Answer Key • Resources) | ||
Preceded by Problem 13 |
Followed by Problem 15 | |
1 • 2 • 3 • 4 • 5 • 6 • 7 • 8 • 9 • 10 • 11 • 12 • 13 • 14 • 15 | ||
All AIME Problems and Solutions |