The sum of absolute values of all real numbers x , such that both of the fractions 7 x 2 − 6 x − 5 x 2 + 4 x − 1 and 1 + x 1 − x are integers, can be written as b a , where a and b are coprime positive integers. What is the value of a + b ?
Details and assumptions
If you think that there are no real numbers x , then the (empty) sum is 0, and you should type your numerical answer as 1.
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.
Great job, straight to the point!
I try to do it with the first fraction and let it be k for some integers k, then I used discriminant, that is, it should be a perfect square to has integer solutions.. But i missed -1/2. Great solution!
Starting with the easier fraction, we know that 1 + x 1 − x = N for some integer N . Solving for x gives x = N + 1 2 − 1 . Since the first fraction must also be integral for the same value of x , it follows that 7 ( N + 1 2 − 1 ) 2 − 6 ( N + 1 2 − 1 ) − 5 ( N + 1 2 − 1 ) 2 + 4 ( N + 1 2 − 1 ) − 1 ∈ Z . Basic algebraic manipulation yields 2 N 2 − 6 N − 1 − N 2 − N + 1 ∈ Z for some integer N .
Now some critical intervals of this function will be tested. Let f ( x ) = 2 x 2 − 6 x − 1 − x 2 − x + 1 To find the asymptote of the function, note that lim x → ± ∞ f ( x ) = − 1 / 2 . From ( − ∞ < x ≤ − 2 , − 1 < y < 0 , so no solutions exist on this interval. From 8 ≤ x < ∞ , − 1 < y < 0 , so no solutions exist on this interval. From − 1 ≤ x ≤ 7 , each individual number will be tested. The results of this are:
x − 1 0 1 2 3 4 5 6 7 ∣ ∣ ∣ ∣ ∣ ∣ ∣ ∣ ∣ ∣ f ( x ) 7 1 − 1 5 1 1 1 1 − 7 1 9 − 1 9 2 9 − 3 5 4 1 − 1
As you can see, integer solutions exist only for N = 0 , 2 , 3 , 7 . From this, we can see that the values of x are N + 1 2 − 1 = 0 + 1 2 − 1 , 2 + 1 2 − 1 , 3 + 1 2 − 1 , 7 + 1 2 − 1 = 1 , − 3 1 , − 2 1 , − 4 3 . Summing the absolute vaules of these yields 1 + 3 1 + 2 1 + 4 3 = 1 2 3 1 . 3 1 + 1 2 = 4 3 .
This is a nice solution, though the statements
"From ( − ∞ < x ≤ − 2 , − 1 < y < 0 " and "From 8 ≤ x < ∞ , − 1 < y < 0 " deserve more explanation.
Most submitted solutions included analysis of the graphs, using increase/decrease and/or continuity, sometimes not totally rigorously. Some solutions involved relatively weak inequalities and an unnecessarily large amount of case-by-case checking.
Let
7 x 2 − 6 x − 5 x 2 + 4 x − 1 ---------eqn A
1 + x 1 − x --------eqn B
Analyzing eqn B,
Let 1 + x 1 − x = p
if p ∈ Z then
∣ 1 + x ∣ ≤ ∣ 1 − x ∣ for some x, x = 1
when x > 1
∣ 1 + x ∣ > ∣ 1 − x ∣
⇒ x ≤ 1 ( Upper bound ) if p must be an integer
Analyzing eqn A,
7 x 2 − 6 x − 5 x 2 + 4 x − 1 = 0 has roots at ( − 5 − 2 , 5 − 2 )
for all x < − 5 − 2 , 0 < e q n ( A ) < 1
proof:
At x < − 5 − 2 ,both x 2 + 4 x − 1 and 7 x 2 − 6 x − 5 are positive
for eqn A to be > 1
∣ x 2 + 4 x − 1 ∣ > ∣ 7 x 2 − 6 x − 5 ∣
as both are positive
x 2 + 4 x − 1 > 7 x 2 − 6 x − 5
6 x 2 − 1 0 x − 4 < 0
− 3 1 < x < 2 which does not tally with x < − 5 − 2
Hence we have a lower bound for x at 5 − 2 since eqn A never evaluates to an integer at x < − 5 − 2
Range : − 5 − 2 ≤ x ≤ 1
However, in this range eqn A and eqn B both have singularities and both can take infinite integer results (not necessarily at the same value of x)
To reduce the range, we know that eqn A = 0 at 5 − 2 . The next integer result that eqn A can take within this range is − 1
7 x 2 − 6 x − 5 x 2 + 4 x − 1 = − 1
x = − 4 3 , x = 1
Range : − 4 3 ≤ x ≤ 1
In this range, eqn B has a continous curve
At x = 1 , e q n B = 0
At x = − 4 3 , e q n B = 7
B | x | A
0 | 1 | -1
1 | 0 | 0.2
2 | -1/3 | 1
3 | -1/2 | 11
4 | -3/5 | -2.714
5 | -2/3 | -1.526
6 | -5/7 | -1.171
7 | -3/4 | -1
1 + 3 1 + 2 1 + 4 3 = 1 2 3 1
b a = 1 2 3 1
a + b = 3 1 + 1 2 = 4 3
"To reduce the range, we know that eqn A = 0 at 5 − 2 . The next integer result that eqn A can take within this range is − 1 " The Intermediate Value Theorem is implicitly used here to show that there are no other integer solutions between 5 − 2 and − 4 3 .
"Range: − 4 3 ≤ x ≤ 1 In this range, eqn B has a continous curve" This is not stated sufficiently strongly. However, B is decreasing at this interval, which is the needed condition.
From 1 + x 1 − x = n obtain x = 1 + n 1 − n where n is integer. Substitute this value of x in the first fraction to obtain that 2 n 2 − 6 n − 1 n 2 + n − 1 must be integer. ∣ n 2 + n − 1 ∣ ≥ ∣ 2 n 2 − 6 n − 1 ∣ is equivalent to ( n 2 + n − 1 ) 2 − ( 2 n 2 − 6 n − 1 ) 2 ≥ 0 which has solutions (for integer n ) the numbers n = 0 , n = 2 , ⋯ , 7 . Checking this numbers we see that the fraction is integer only when n = 0 , n = 2 , n = 3 , n = 7 . Now calculate x = 1 + n 1 − n and add their absolute values to get 1 + 3 1 + 2 1 + 4 3 = 1 2 3 1 .
Let k ( 1 + x ) = 1 − x . Then, solving, we have k + k x = 1 − x , ( k + 1 ) x = 1 − k , x = k + 1 1 − k . Letting j = k + 1 gives the simpler x = j 2 − j . Let x ( j ) = j 2 − j . Then x ′ ( j ) = j 2 ( 2 − j ) ′ ∗ j − ( j ) ′ ∗ ( 2 − j ) = j 2 ( − 1 ) ∗ j − ( 1 ) ∗ ( 2 − j ) = j 2 − j − 2 + j = − j 2 2 . Therefore x ′ is always negative, and x decreases as j increases. Splitting into two cases, we have when j > 0 that x ≤ 1 2 − 1 = 1 , and when j < 0 , x ≥ − 1 2 − ( − 1 ) = − 1 3 = − 3 . Also, j 2 − j + 1 = j 2 − j + j = j 2 , which is < 0 when j < 0 (meaning that j 2 − j + 1 < 0 and j 2 − j < − 1 , so x < − 1 ), and is > 0 when j > 0 (meaning, similarly, that x > − 1 ). Combining all together (and neglecting x = 1 ), we have the nice simple inequality − 3 ≤ x ≤ 1 . Now, we know by the Mean Value Theorem that if a function is not monotone on an interval, either a) its derivative is 0 somewhere in the interval, or b) it is discontinuous somewhere in the interval. We will check both for the fraction 7 x 2 − 6 x − 5 x 2 + 4 x − 1 . a) Let the fraction be f ( x ) to simplify notation. Then f ′ ( x ) = ( 7 x 2 − 6 x − 5 ) 2 ( x 2 + 4 x − 1 ) ′ ∗ ( 7 x 2 − 6 x − 5 ) − ( 7 x 2 − 6 x − 5 ) ′ ∗ ( x 2 + 4 x − 1 ) = ( 7 x 2 − 6 x − 5 ) 2 ( 2 x + 4 ) ( 7 x 2 − 6 x − 5 ) − ( 1 4 x − 6 ) ( x 2 + 4 x − 1 ) , which we shall check for zeros. Setting it equal to 0 and multiplying away the denominator gives us ( 2 x + 4 ) ( 7 x 2 − 6 x − 5 ) − ( 1 4 x − 6 ) ( x 2 + 4 x − 1 ) = 0 . Expanding the LHS gives 1 4 x 3 + 1 6 x 2 − 3 4 x − 2 0 − ( 1 4 x 3 + 5 0 x 2 − 3 8 x + 6 ) = − 3 4 x 2 + 4 x − 2 6 = 0 . The discriminant of this quadratic is 1 6 − 4 ∗ 2 6 ∗ 3 4 ≪ 0 , so this equation has no roots. b) A rational function is discontinuous only when its denominator is 0, so we solve 7 x 2 − 6 x − 5 = 0 to yield the solutions 7 3 ± 2 1 1 , which are approximately − 0 . 5 2 and 1 . 3 8 . Therefore f ( x ) is monotone in two distinct subintervals of [-3,1]; [-3, -0.52...] and [-0.52..., 1]. Now we just plug in values for j and see where x and f(x) are! We only care about places where integer values are possible. We already know that the first interval applies for j < 0 (since all x values there are < − 1 < − 0 . 5 2 . . . . Checking j > 0 , we find that the only values in the second interval are where j = 1 , 2 , 3 , 4 (In this case, we find that x = 1 , 0 , − 3 1 , − 2 1 ). In the first interval, when j < 0 we have f ( − 3 ) = − 1 9 1 ≥ f ( x ) > f ( − 1 ) = − 2 1 , an interval with no integer values. When j > 4 , the values of x are between x ( 5 ) = − 5 3 and − 1 , so f ( − 5 3 ) = − 7 1 9 ≤ f ( x ) < − 2 1 = f ( − 1 ) . This interval has two integer values in it, and although -2 is not reachable (I won't show you, but there are two consecutive values straddling -2), -1 certainly is when j = 8 and x = − 4 3 . Therefore − 4 3 is a potential value of x. Quickly checking the four values of x corresponding to j = 1 , 2 , 3 , 4 , we find surprisingly that 3 of them work: f ( 1 ) = − 1 , f ( − 3 1 ) = 1 , and f ( − 2 1 ) = 1 1 . We have 4 values of x, and have rigorously shown that these are the only 4 values. Therefore the sum of their absolute values is 4 3 + 1 + 3 1 + 2 1 = 1 2 3 1 , and our answer is 3 1 + 1 2 = 4 3 .
This solution, though long, relies on one simple, intuitive fact: Most functions are monotone most of the time!
First, we take 1 + x 1 − x = b (b is an integer) → 1 + b 1 − b = x . We then replace x in the first factor and receive another factor: 1 + 6 b − 2 b 2 b 2 + b − 1 = a ( a is an integer) → ( 2 a + 1 ) b 2 + ( 1 − 6 a ) b − 1 − a = 0 . This equation has two roots . Now we find a for the first root b = 2 ( 2 a + 1 ) 6 a − 1 + 4 4 a 2 + 5 is integer. Note that 2 ( 2 a + 1 ) 6 a − 1 + 4 4 a 2 + 5 = 2 3 + 2 ( 2 a + 1 ) 4 4 a 2 + 5 − 4 . We can infer that 2a+1 must divides 4 4 a 2 + 5 . If ∣ a ∣ > 1 1 we have the inequality : 4 ( 2 a + 1 ) > 4 4 a 2 + 5 − 4 > 3 ( 2 a + 1 ) , a contradiction so ∣ a ∣ ≤ 1 1 . Just calculate and we will have a = ± 1 and a = 1 1 that give x = 3 − 1 , 2 − 1 , 1 , 4 − 3 . The another roots of b is just the same but it has no integer a that satisfy b is integer.
Note : m^n denotes m to the power of n
Let f(x) = ( x^2 + 4 * x - 1 ) / ( 7 * x^2 - 6 * x - 5 )
and g(x) = ( 1 - x / 1 + x )
Claim : g(x) is not a rational number for any irrational value of x.
Proof : Let there exists an irrational no. x0 such that g(x0) is equal to an
integer say r. Then
r = ( 1 - x0 ) / ( 1 + x0)
which implies
x0 = ( 1 - r ) / ( 1 + r )
which is rational since r is rational.
This implies x0 is rational which is a contradiction to our assumption.
Hence, g(x) cannot be rational and therefore an integer for any irrational value of x.
Now, g(1) = 0 and f(1) = -1 , so x = 1 is a solution.
Now, modulus of any non-zero integer is greater than 1.
So, if g(x) is an integer, then
| 1 - x | ≥ | 1 + x |
From which we get either x < -1 or -1 < x ≤ 0
Let x = m / n where n is a natural number and m is an integer such that
gcd ( m , n ) = 1
Also since | x | < 1 it implies that | m | < n
Then g(x) = ( n - m ) / ( n + m ) = 1 - ( 2 * m / m + n )
This will be an integer in the following cases
i) m + n = 1
but n ≥ 1 implies m ≤ 0
So putting x = ( 1 - n ) / n we get g ( x ) = 2 * n - 1 for natural no. n
ii) m + n = -1 but n ≥ | m | implies m + n ≥ 0, so this case is invalid
iii) m + n = 2 but n ≥ 1 implies m ≤ 1
So putting x = ( 2 - n ) / n we get g (x ) = n - 1 for natural no. n
iv) m + n = -2 which is again invalid since m + n ≥ 0
v) m + n = 2 * m i.e. m = n i.e. x = 1
vi) m + n = -2 * m i.e. n = -3 * m
or x = - 1 / 3 and g( x ) = 2
but this solution is included in the solution of case ( iii ) i.e. m + n = 2
Now, of the above values of x we have to find those which make f(x) an integer.
i) x = 1 for which g ( x ) = 0 and f ( x ) = -1
ii) x = ( 1 - n ) / n for natural n or putting 1 - n = -k for whole no. k i.e. x = -k / ( 1 + k ) we get
f ( x ) = - ( 4 * k ^2 + 6 * k + 1 ) / ( 8 * k^2 - 4 * k - 5 )
For f(x) to be an integer | f ( x ) | ≥ 1 ( since f ( x ) ≠ 0 for any x )
|- ( 4 * k ^2 + 6 * k + 1 ) | ≥ | ( 8 * k^2 - 4 * k - 5 ) |
But 4 * k ^2 + 6 * k + 1 > 0 for all natural k and
8 * k^2 - 4 * k - 5 > 0 for all natural k > 1
So, for k = 1 we get x = -1 / 2 and f (x) = 11
and for k > 1 we get
9 * k^2 - 5 * k - 14 ≤ 0
This means 1 < k ≤ 3 but for k = 2 ,
f(x) is not an integer and for k = 3
x = -3 / 4 and f(x) = -1
iii) x = ( 2 - n ) / n for natural n or putting 2 - n = -k for integer k
i.e. x = -k / ( 2 + k ) we get
f ( x ) = - 4 * ( k ^2 + 3 * k + 1 ) / ( 13 * k^2 + 7 * k - 10 )
For f(x) to be an integer | f ( x ) | ≥ 1 which means
| - 4 * ( k ^2 + 3 * k + 1 ) | ≥ | ( 13 * k^2 + 7 * k - 10 ) |
Solving similarly as above we get two values of k which are
k = -1 i.e. x = 1 for which f(x) = -1 and
k = 1 i.e. x = -1 / 3 for which f(x) = 1
Hence , all the values of x for which f(x) and g(x) are integers are
x = 1 ,- 1 / 2 , - 1 / 3, - 3 / 4
The sum of absolute values of the above values of x = 31 / 12.
Since gcd ( 31 , 12 ) = 1, a = 31 and b =12.
Hence a + b = 43 which is the required answer.
1 + x 1 − x = 1 + x 2 − 1 ∈ Z ⇒ 1 + x 2 = k ∈ Z ⇒ x = k 2 − 1
Substituting this into the first fraction, 7 x 2 − 6 x − 5 x 2 + 4 x − 1 = 7 ( k 2 − 1 ) 2 − 6 ( k 2 − 1 ) − 5 ( k 2 − 1 ) 2 + 4 ( k 2 − 1 ) − 1 × k 2 k 2 = \frac{-4k^2+4k+4}{8k^2-40k+28} = \frac{-k^2+k+1}{2k^2-10k+7})
My solution is not strictly under Mathematics I first converted these two fractions in one fraction of k(which is integer) which yields integer
Then i run loop of k from 0 to 100 to gain values of k and x
The conversion is very useful. Instead of running the loop, one can complete the problem mathematically. Does anybody know how to do that?
Log in to reply
Only k = 0 , . . . , 7 need to be checked -- see my explanation below Jiaqi W's solution.
Log in to reply
Yes, this is, probably, the easiest way to do it. One can also play with the g.c.d.
Thanks this is best
Let 1 + x 1 − x = a and 7 x 2 − 6 x − 5 x 2 + 4 x − 1 = b with a and b both being integers. Then 1 − x = a + a x ⇒ x = 1 + a 1 − a
Plug x = 1 + a 1 − a into 7 x 2 − 6 x − 5 x 2 + 4 x − 1 = b , we get 2 a a − 6 a − 1 1 − a − a a = b
1 − a − a 2 = ( 2 b ) a 2 + ( − 6 b ) a + ( − b ) ⇒ 0 = ( 2 b + 1 ) a 2 + ( 1 − 6 b ) a + ( − b − 1 )
Now, use the quadratic formula. Since a is an integer, the discriminant must be a perfect square:
( 1 − 6 b ) 2 − 4 ( 2 b + 1 ) ( − b − 1 ) ⇒ 4 4 b 2 + 5 = 0
a = 4 b + 1 ( 6 b − 1 ) ± 4 4 b 2 + 5
After trying a few values of b, we find that for b = ± 1 the discriminant is 7, and for b = ± 1 1 the discriminant is 73.
Now, we plug b back into the equation a = 4 b + 1 ( 6 b − 1 ) ± 4 4 b 2 + 5 :
b = 1 a = 2 , 3 − 1
b = − 1 a = 0 , 7
b = 1 1 a = 3 , 2 3 − 4
b = − 1 1 a = 7 − 1 , 3 1 0
Taking only the integer solutions, a = 0 , 2 , 3 , 7 . Plug this back into x = 1 + a 1 − a .
x = 2 − 1 , 3 − 1 , 4 − 3 , 1
Adding the absolute value of each x gives 1 2 3 1 , so the answer is 43 .
This solution probably cannot be completed, but the comments raise very interesting questions.
I got this far (although failed at computing 31+12). However there are other solutions to 4 4 b 2 + 5 being a square: 4 4 ⋅ 4 0 9 2 + 5 = 2 7 1 3 2 , 4 4 ⋅ 4 3 7 9 2 + 5 = 2 9 0 4 7 2 . There may be others (I found these with a computer). However none of these extra values gave an integer for the second equation.
So nobody has posted a compete solution yet? Is the problem solvable?
For the 4 4 b 2 + 5 being a square part, i just tried some values for b. There should be a better way though.
Log in to reply
I'm not sure either, but we at least know that b must be odd:
4 4 b 2 + 5 clearly must be odd, thus it must equal an odd perfect square.
Odd perfect squares = ( 2 n + 1 ) 2
4 ( n ( n + 1 ) ) + 1 = 4 4 b 2 + 5
n ( n + 1 ) = 1 1 b 2 + 1 .
Since n(n+1) is even, b must be odd. Thus, b 2 = 4 n + 1 → 4 4 n + 1 2 = n ( n + 1 ) so we also know that n or n + 1 is a multiple of 4 .
Anyway, what I'm more interested in is evidence that 1 and 1 1 are the only b which satisfy the equation.
So it's clear that for b = ± 1 the discriminant is 7 2 , and for b = ± 1 1 the discriminant is 7 3 2 . But then you're stuck in terms of seeing whether there are any other possibilities.
Here's how I did it: we can show that b 2 ≥ 1 implies that a ( a − 7 ) ( a − 2 ) ( 3 a + 1 ) ≤ 0 , which implies − 3 1 ≤ a ≤ 7 . Then we have only 8 integer a-values to test (along with b = 0 which is easily ruled out). I didn't use the quadratic formula or solve the equation for a at all.
Log in to reply
can you explain the second paragraph in more detail?
Log in to reply
@Matt McNabb – Basically, see what you get when you plug the formula for b into b 2 ≥ 1 ... or, if you prefer, into ∣ b ∣ ≥ 1 .
You wrote a a instead of a 2 in your third line.
I think there are infinitely many b such that 4 4 b 2 + 5 is perfect square, but only those that you found yield integral a . Try e.g. b = 6 4 7 8 6 4 2 9 ;D
I lacked -1/2 But this is interesting because the solution is the most clearly stated than others do. Thanks
Discriminant is 7 2 and not 7.
Analisemos a segunda fracção, uma vez que é menos 'complexa'!!
7 x 2 − 6 x − 5 x 2 + 4 x − 1 = − 1 4 x 2 − x − 3 = 0 ( x − 1 ) ( x + 4 3 ) = 0
Aplicando raciocínio análogo, obtemos: 1, - 3/4, - 1/2 e - 1/3.
Portanto,
∣ 1 ∣ + ∣ ∣ − 4 3 ∣ ∣ + ∣ ∣ − 2 1 ∣ ∣ + ∣ ∣ − 3 1 ∣ ∣ = 1 + 4 3 + 2 1 + 3 1 = 1 2 3 1
Daí,
a + b = 3 1 + 1 2 a + b = 4 3
English, please!
Problem Loading...
Note Loading...
Set Loading...
Let k be an integer such that 1 + x 1 − x = k .Then we have x = 1 + k 1 − k .Substitute this to the first fraction,we have the fraction 2 k 2 − 6 k − 1 − k 2 − k + 1 .For this to be an integer,we must have ∣ 2 k 2 − 6 k − 1 ∣ ≤ ∣ − k 2 − k + 1 ∣ ⇒ ( k 2 − 7 k ) ( 3 k 2 − 5 k − 2 ) ≤ 0 ⇒ k ( k − 7 ) ( 3 k + 1 ) ( k − 2 ) ≤ 0 which have solutions − 3 1 ≤ k ≤ 0 and 2 ≤ k ≤ 7 .Substituting the possibilities of k ,we have k = 0 , 2 , 3 , 7 satisfying the requirement,with respective values of x = 1 , − 3 1 , − 2 1 , − 4 3 .Hence,the sum of the absolute values of x is 1 + 3 1 + 2 1 + 4 3 = 1 2 3 1 ,which yields the answer 3 1 + 1 2 = 4 3