ProbXiv
Explorersign in
Problem archiveProblem record

Statement

Let P1P_1 and P2P_2 be uniform distributions on closed intervals [a,b][a,b] and [c,d][c,d], respectively, where a<b<c<da<b<c<d. Let P:=pP1+(1−p)P2P:=pP_1+(1-p)P_2 be the mixture associated with the probability vector (p,1−p)(p,1-p), where 0<p<10<p<1. For each nn, let αn\alpha_n be an optimal nn-point set for PP, and let Vn(P)V_n(P) denote its quantization error.

Asymptotic separation conjecture. For every probability vector (p,1−p)(p,1-p), there exists a positive integer NN such that, for all negNn eg N, the optimal set αn\alpha_n contains points from both [a,b][a,b] and [c,d][c,d], contains no point from (b,c)(b,c), and, if k:=card⁡(αn∩[a,b])k:=\operatorname{card}(\alpha_n\cap[a,b]), then

αn(P)=αk(P1)∪αn−k(P2),Vn(P)=1108(pk2+1−p(n−k)2).\alpha_n(P)=\alpha_k(P_1)\cup\alpha_{n-k}(P_2),\qquad V_n(P)=\frac{1}{108}\left(\frac{p}{k^2}+\frac{1-p}{(n-k)^2}\right).

The conjecture predicts that sufficiently many quantization points eventually lie in the two supporting intervals and split into optimal sets for the two component distributions. It would settle the corresponding asymptotic behavior for mixtures of uniform distributions on separated intervals, while the source presents the related low-rate placement questions as open.

Record

Comments

Nothing has been published against this problem yet, and nothing has been said about it.

Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.

Sign in

Solve with an agent

Open the statement in a chat, with the problem and the ground rules already written into the prompt.

This opens a third-party site. Nothing is posted back to ProbXiv and nothing you write there is recorded here — what a model gives you is an attempt, which a person still has to check.