Partial homogenization conjecture for adaptive weight systems

Consider the SDE system with coefficients satisfying Assumption~ and ϵ=1\epsilon=1. For kNk\in\mathbb{N}, let deterministic initial data (Xi(k)(0))i[Nk](X^{(k)}_i(0))_{i\in[N_k]} and (wi,j(k)(0))i,j[Nk](w^{(k)}_{i,j}(0))_{i,j\in[N_k]} satisfy Assumption~, and suppose

limksupk1,k2kγW1,(((Xi(k1)(0))i[Nk1],(wi,j(k1)(0))i,j[Nk1],[Nk1]),((Xi(k2)(0))i[Nk2],(wi,j(k2)(0))i,j[Nk2],[Nk2]))=0.\lim_{k\to\infty}\sup_{k_1,k_2\geq k}\gamma_{W_1,\square}\bigg(\Big((X^{(k_1)}_i(0))_{i\in[N_{k_1}]},(w^{(k_1)}_{i,j}(0))_{i,j\in[N_{k_1}]},[N_{k_1}]\Big),\Big((X^{(k_2)}_i(0))_{i\in[N_{k_2}]},(w^{(k_2)}_{i,j}(0))_{i,j\in[N_{k_2}]},[N_{k_2}]\Big)\bigg)=0.

Partial homogenization conjecture. If the corresponding solutions of are denoted by (Xi(k)(t))i[Nk](\bm{X}^{(k)}_i(t))_{i\in[N_k]} and (wi,j(k)(t))i,j[Nk](\bm{w}^{(k)}_{i,j}(t))_{i,j\in[N_k]}, then

limtlim supksupk1,k2kE[δW1,(((Xi(k1)(t))i[Nk1],(wi,j(k1)(t))i,j[Nk1],[Nk1]),((Xi(k2)(t))i[Nk2],(wi,j(k2)(t))i,j[Nk2],[Nk2]))]=0.\lim_{t\to\infty}\limsup_{k\to\infty}\sup_{k_1,k_2\geq k}\mathbb{E}\bigg[\delta_{W_1,\square}\bigg(\Big((\bm{X}^{(k_1)}_i(t))_{i\in[N_{k_1}]},(\bm{w}^{(k_1)}_{i,j}(t))_{i,j\in[N_{k_1}]},[N_{k_1}]\Big),\Big((\bm{X}^{(k_2)}_i(t))_{i\in[N_{k_2}]},(\bm{w}^{(k_2)}_{i,j}(t))_{i,j\in[N_{k_2}]},[N_{k_2}]\Big)\bigg)\bigg]=0.

The claim predicts that initial data asymptotically indistinguishable in the weaker metric become indistinguishable in the stronger metric at long times, expressing partial homogenization of the weight structures.

Sources & referencesView supporting material

Primary source

Datong Zhou, “Non-exchangeable mean-field theory for adaptive weights: propagation of dissociatedness and graphon sampling lemma”, arXiv:2506.13587 (2025).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.