Given a positive integer k>1 and n distinct points in k-dimensional coordinates a(1)=(a1(1),…,ak(1)), …,a(n)=(a1(n),…,ak(n)), we define the "score" of a(i) as follows: j=1∏k#{i′∣1≤i′≤n such that πj(a(i′))=πj(a(i))}, where #S denotes the number of elements in the set S, and πj:Rk→Rk−1 is the projection map that sends a point in k-dimensional coordinates to the point obtained by deleting its j-th coordinate. Find the largest real number t such that the t-th power mean of the "scores" of all a(i) is ≤n.
Note: The t-th power mean of positive real numbers x1,…,xn is defined as ( x 1 t + + x_n^t}{n} )t1(if } t = 0); nx1⋯ x_n} (if } t = 0).
For every positive integer k greater than 1, find the largest real number t such that the following hold: Given n distinct points a(1)=(a1(1),⋯,ak(1)), ⋯, a(n)=(a1(n),⋯,ak(n)) in Rk, we define the score of the tuple a(i) as j=1∏k#{1≤i′≤n such that πj(a(i′))=πj(a(i))} where #S is the number of elements in set S, and πj is the projection Rk→Rk−1 omitting the j-th coordinate. Then the t-th power mean of the scores of all a(i)'s is at most n.
Note: The t-th power mean of positive real numbers x1,⋯,xn is defined as (nx1t+⋯+xnt)t1
Solution
The answer is k−11. To see that t≤k−11, consider the sequences where all but one of the components are equal to 1, and the remaining one component is 1,2,…,m. Then n=km+1 and the score of the zero sequence is mk. In order for the statement to hold, we must have (km+1mkt)t1≤km+1, i.e., mk≤(km+1)1+1/t. By choosing m to be sufficiently large, we get k≤1+t1, or t≤k−11.
Now we will prove by induction on k that t=k−11 works. We first assume that k≥3 and the statement holds for k−1. For every real number r, let cr be the number of sequences a(i) with a1(i)=r. For each r with cr=0, we gather the sequences with a1(i)=r and compute their scores when we remove the first component. Let the scores be s1′,⋯,scr′. Moreover, for each truncated sequence, we see how many sequences from the original list would have produced the same truncated sequence. Let the number of such sequences be d1,⋯,dcr. Then clearly d1+⋯+dcr≤n as all sequences are distinct. Note that the scores of the original sequences we gather are precisely d1s1′,⋯,dcrscr′. By the inductive hypothesis, we have s1k−21+⋯+scrk−21=crk−2k−1. By Hölder's inequality, we have ((d1s1′)k−11+⋯+(dcrscr′)k−11)k−1≤(d1+⋯+dcr)⋅(s1k−21+⋯+scrk−21)k−2≤ncrk−1. By taking the (k−1)-th root of both sides and adding it up for all r∈R with cr=0, we get that the k−11-th power mean is bounded by (n∑cr=0nk−11cr)k−1=n as cr=0∑cr=n. For the base case k=2, we can follow the same proof except that we clearly have s1′=⋯=scr′=n and so (d1s1′+⋯+dcrscr′)=n(d1+⋯+dcr)≤n2 and the rest of the proof follows verbatim.
Want a route through all this instead of an archive? The track
puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.
Source: MathNet,
licensed CC-BY-4.0.
Statement translated into English from zh; metadata (topic, difficulty) added by this project.