Regd.

Office : Aakash Tower, 8, Pusa Road, New Delhi-110005

Ph.: 011-47623456

INMO-2019 WORKSHOP
HOME TEST-2

SOLUTION

1. Since b is the product of roots of the quadratics, of the form 3k or 3k + 1, where k is any integer.
it must be expressible as a product of integers in Thus there are infinitely many pairs (a, b), given
two different ways. So let us try for b numbers of by (i), where n + 1 is not divisible by 3, satisfying
the form uvw. Suppose that the given conditions.

u + vw = a, uv + w = 2a If n + 1 is divisible by 3, and a, b are given by (i)

then a is divisible by 3 and b is divisible by 9. It
Then 2u + 2vw = uv + w; w(2v – 1) = u(v – 2).
a b
This will be satisfied if u = 2v – 1, w = v – 2 can be seen that  ,  will satisfy our
3 9
where v is any integer. If v = n this gives
conditions. Clearly, if (a, b) satisfies the
a = (2n – 1) + n(n – 2) = n2 – 1, conditions, then (–a, b) also will satisfy the
b = n(n – 2)(2n – 1) ...(i) conditions.

If a and b are chosen like this, then the roots of 2. If p is prime and p 2 then

  
the first equation are p 1 p 1

–u = –(2n – 1), –vw = –n(n – 2); 2 p 1  1  2 2  1 2 2 1

those of the second equation are is divisible by p, by Fermat’s theorem. So either
–uv = –n(2n – 1), –w = –(n – 2). p 1 p 1
2 2  1 or 2 2  1 is divisible by p. They are
These are integers. So if a and b are relatively
prime then the conditions are fulfilled. Let p be relatively prime since they are successive odd
any prime number dividing a = (n – 1)(n + 1). If integers. Suppose that p is an odd prime such
p divides n – 1, then p does not divide any of the 2p  1
numbers: that is a square. First suppose that
p
n – 1 + 1 = n, n – 1 – 1 = n – 2, 2(n – 1) + 1 = p 1
2n – 1; 2 2  1 is divisible by p. We know that if x and
hence p does not divide b. Suppose that p divides y are relatively prime positive integers such that
n + 1. then p does not divide their product xy is a perfect square then x and y
are separately perfect squares.
n + 1 – 1 = n, n + 1 – 3 = n – 2, 2(n + 1) – 3 =
2n – 1,

 
 p 1  p 1
provided p 3. If n + 1 is not divisible by 3 this 2p  1  2 2  1 
requirement also is satisfied. So a and b cannot Since   2
2 1 ,
p  p 
have a common prime divisor if n is any integer

(1)
 p 1
a 1 (mod 5) b 1 (mod 5); a 2 (mod 5)
we see that 2 2  2
1  a for some integer a > 2. b 3 (mod 5);
a 3 (mod 5) b 2 (mod 5); a 4 (mod 5)
Thus
b 4 (mod 5);
p 1 pa3 + qa2 + ra + s 0 (mod 5)
2 2  a  1   a  1 a  1 ;
2
(pa3 + qa2 + ra + s)b3 0 (mod 5)
If p(ab)3 + q(ab)2b + r(ab)b2 + sb3 0 (mod 5)
a–1= 2r, a+1= 2s, 0 < r < s, p + qb + rb2 + sb3 0 (mod 5)
2s – 2r = 2, 2s – 1 – 2r – 1 = 1. Hence, for each a, setting b to be its multiplicative
Since 0 r – 1 < s – 1 this is possible only if r – inverse (mod 5) is our desired b.
1 = 0, s – 1 = 1. 4. We seek solutions (x, y, z) which are in arithmetic
Hence progression. Let us put y – x = z – y = d > 0 so
that the equation reduces to the form
p 1
p 1 3y2 + 2d2 = 2d3.
a  1  2, a  3, 2 2  a 2  1  8,  3, p  7
2 Thus we get 3y2 = 2(d – 1)d2. We conclude that
2(d – 1) is 3 times a square. This is satisfied if
p 1
d – 1 = 6n2 for some n. Thus d = 6n2 + 1 and 3y2
Next suppose that 2 2  1 is divisible by p.
= d2. 2(6n2) giving us y2 = 4d2n2. Thus we can
Since take y = 2dn = 2n(6n2 + 1). From this we obtain
x = y – d = (2n – 1)(6n2 + 1), z = y + d = (2n +
1)(6n2 + 1). It is easily verified that
 
p 1  p 1 
2p  1  2 2  1
, 2 2  1  (x, y, z) = ((2n – 1)(6n2 + 1), 2n(6n2 + 1), (2n +
p  p  1)(6n2 + 1)),
is indeed a solution for a fixed n and this gives an
p 1
we see, exactly as before, that 2 2  1 b 2 for infinite set of solutions as n varies over natural
numbers.
some integer b > 0.
5. We show that the answer is NO. Suppose, if
p 1 possible, let a, b, c be three distinct positive real
If  2 then this implies that b2 –1 (mod 4) numbers such that a, b, c, b + c – a, c + a – b,
2
a + b – c and a + b + c form a 7-term arithmetic
p 1 progression in some order. We may assume that
which is not possible. Hence  1; p  3, p
2 a < b < c. Then there are only two cases we need
being odd. to check: (I) a + b – c < a < c + a – b < b < c <
b + c – a < a + b + c and (II) a + b – c < a < b <
 2p  1  1  9 c + a – b < c < b + c – a < a + b + c.
Finally, if p = 7 then , and if p = 3
p Case I. Suppose the chain of inequalities a + b –
c<a<c+a–b<b<c<b+c–a<a+b+c

then
 2p  1  1  1 . Thus the desired primes are holds good. Let d be the common difference.
p Thus we see that

3 and 7. c = a + b + c – 2d, b = a + b + c – 3d, a = a + b

+ c – 5d.
3. Note that if 5 | a, then 5 | pa3 + qa2 + ra and since
Additing these, we see that a + b + c = 5d. But
a 3 + qa 2 + ra + s.
5 s , we obtain that 5 pa
then a = 0 contradicting the positively of a.
Hence, 5 a . This means a ±1, ±2 (mod 5). For
Case II. Suppose the inequalities a + b – c < a
every such a, there exists b z, such that ab
< b < c + a – b < c < b + c – a < a + b + c are
1 (mod 5).
true. Again we see that

(2)
 c = a + b + c – 2d, b = a + b + c – 4d, a = a + b p(pn – 1 + pn – 2 + ... + 1) = q(q + 1).
+ c – 5d.
If q pn/2 – 1, then q < pn/2 and hence we see
that q2 < pn. Thus we obtain
 11 
We thus obtain a + b + c =   d . This gives
2 q2 + q < pn + pn/2 < pn + pn – 1 + ... + p,
since n > 2. It follows that q pn/2. Since n > 2
1 3 7 and is an even number, n/2 is a natural number
a d , b  d, c  d .
2 2 2 larger than 1. This implies that q pn/2 by the
given condition that q is a prime. We conclude
 1 that q pn/2 + 1. We may also write the above
Note that a + b – c = a + b + c – 6d = –   d .
2 relation in the form
However we also get a + b – c p(pn/2 – 1)(pn/2 + 1) = (p – 1)q(q + 1).
This shows that q divides (pn/2 – 1)(pn/2 + 1). But
 1   3      3
=          d     d . It fllows that 3e q pn/2 + 1 and q is a prime. Hence the only
 2   2   2   2
possibility is q = pn/2 + 1. This gives
= e giving d = 0. But this is impossible.
p(pn/2 – 1) = (p – 1)(q + 1) = (p – 1)(pn/2 + 2).
Thus there are no three distinct positive real
Simiplification leads to 3p = pn/2 + 2. This shows
numbers a, b, c such that a, b, c, b + c – a, c +
that p divides 2. Thus p = 2 and hence q = 5,
a – b, a + b – c and a + b + c form a 7-term
n = 4. It is easy to verify that these indeed satisfy
arithmetic progression in some order.
the given equation.
6. Obviously p q. We write this in the form



(3)