Browse · MathNet
PrintIreland
Ireland number theory
Problem
Find all pairs of integers such that .
Solution
First Solution: The key to this solution is the identity Even though this can be verified readily, we explain how to find it. We start with a polynomial of degree 3 which has , and an auxiliary number as its roots . We obtain the identity Substituting gives the identity mentioned above.
The original equation is equivalent to the equation Because we are looking for integer solutions, the two numbers and have to be divisors of . The integer divisors of are . In particular, . We have and . The last identity and imply . Hence, the following table covers all possibilities for and .
From and we obtain by solving the quadratic equation and using . For to be an integer, we need therefore that the discriminant of the quadratic, , is a perfect square. Only the first and third row of the table give . Therefore, the only possible solutions are . Substituting them in the original equation shows that they are indeed solutions.
Second Solution: Because is not a perfect cube, neither nor could be zero. Without loss of generality, we may assume . We distinguish the following three cases.
: In this case, , hence . This implies . If , we get . As , the only possibilities for are but none of them leads to a solution. If , we have to solve . As , can only be or . Only is a solution.
: In this case, . Because and this implies . Hence, . On the other hand, and so which implies . If , we get . As this implies , i.e. which does not give a solution. If , we get . As before, this implies , i.e. or . But both values do not solve the equation.
a < 0 < b: In this case, and so which implies . Let and with . The given equation becomes The left hand side is at least , hence we need to have which implies . We cannot have , because is not a perfect cube. If we need to solve , or equivalently, with as its only positive solution. This leads to . Finally, if we need to have which has no solution as is not divisible by .
Therefore, the only solutions to the original equation are and .
The original equation is equivalent to the equation Because we are looking for integer solutions, the two numbers and have to be divisors of . The integer divisors of are . In particular, . We have and . The last identity and imply . Hence, the following table covers all possibilities for and .
| u | v | a+b | 3ab | ab |
|---|---|---|---|---|
| 1 | 52 | 2 | -45 | -15 |
| -2 | -26 | -1 | 27 | 9 |
| 4 | 13 | 5 | 18 | 6 |
| 13 | 4 | 14 | 207 | 69 |
| -26 | -2 | -25 | 603 | 201 |
| 52 | 1 | 53 | 2862 | 954 |
Second Solution: Because is not a perfect cube, neither nor could be zero. Without loss of generality, we may assume . We distinguish the following three cases.
: In this case, , hence . This implies . If , we get . As , the only possibilities for are but none of them leads to a solution. If , we have to solve . As , can only be or . Only is a solution.
: In this case, . Because and this implies . Hence, . On the other hand, and so which implies . If , we get . As this implies , i.e. which does not give a solution. If , we get . As before, this implies , i.e. or . But both values do not solve the equation.
a < 0 < b: In this case, and so which implies . Let and with . The given equation becomes The left hand side is at least , hence we need to have which implies . We cannot have , because is not a perfect cube. If we need to solve , or equivalently, with as its only positive solution. This leads to . Finally, if we need to have which has no solution as is not divisible by .
Therefore, the only solutions to the original equation are and .
Final answer
(-3, 5), (5, -3), (2, 3), (3, 2)
Techniques
Factorization techniquesTechniques: modulo, size analysis, order analysis, inequalitiesPolynomial operationsVieta's formulas