Browse · MathNet
PrintBMO 2017
2017 geometry
Problem
What is the least positive integer such that, in every convex -gon, the sum of any diagonals is greater than or equal to the sum of the remaining diagonals?
Solution
Let . Consider a convex -gon such that one of its vertices is at and the remaining vertices are within of where is an arbitrarily small positive real. Let equal the total number of diagonals. When , the sum of the shortest diagonals is arbitrarily small. When , the sum of the shortest diagonals is arbitrarily close to and the sum of the remaining diagonals is arbitrarily close to . Therefore, we need to have and .
We proceed to show that works. To this end, colour all remaining diagonals green. To each green diagonal , apart from, possibly, the last one, we will assign two red diagonals and so that no green diagonal is ever coloured red and no diagonal is coloured red twice.
Suppose that we have already done this for green diagonals (thus forming red-red-green triangles) and let be up next. Let be the set of all diagonals emanating from or and distinct from : we have . Every red-red-green triangle formed thus far has at most two sides in . Therefore, the subset of all as-of-yet-uncoloured diagonals in contains at least elements.
When , . The total number of endpoints distinct from and of diagonals in , however, is . Therefore, two diagonals in have a common endpoint and we can assign and to , as needed.
The case is slightly more tricky: this time, it is possible that no two diagonals in have a common endpoint other than and , but, if so, then there are two diagonals in that intersect in a point interior to both. Otherwise, at least one (say, ) of the two vertices adjacent to is cut off from by the diagonals emanating from and at least one (say, ) of the two vertices adjacent to is cut off from by the diagonals emanating from (and ). This leaves us with at most suitable endpoints and at least diagonals in , a contradiction.
By the triangle inequality, this completes the solution.
We proceed to show that works. To this end, colour all remaining diagonals green. To each green diagonal , apart from, possibly, the last one, we will assign two red diagonals and so that no green diagonal is ever coloured red and no diagonal is coloured red twice.
Suppose that we have already done this for green diagonals (thus forming red-red-green triangles) and let be up next. Let be the set of all diagonals emanating from or and distinct from : we have . Every red-red-green triangle formed thus far has at most two sides in . Therefore, the subset of all as-of-yet-uncoloured diagonals in contains at least elements.
When , . The total number of endpoints distinct from and of diagonals in , however, is . Therefore, two diagonals in have a common endpoint and we can assign and to , as needed.
The case is slightly more tricky: this time, it is possible that no two diagonals in have a common endpoint other than and , but, if so, then there are two diagonals in that intersect in a point interior to both. Otherwise, at least one (say, ) of the two vertices adjacent to is cut off from by the diagonals emanating from and at least one (say, ) of the two vertices adjacent to is cut off from by the diagonals emanating from (and ). This leaves us with at most suitable endpoints and at least diagonals in , a contradiction.
By the triangle inequality, this completes the solution.
Final answer
4900
Techniques
Triangle inequalitiesPigeonhole principleColoring schemes, extremal arguments