Conjecture on maximum divergence from trapezoid models

Let \Ta,b,d\T_{a,b,d} be the trapezoid model in the simplex described above, with parameters aa, bb, and dd, and let D(\Ta,b,d)D(\T_{a,b,d}) denote its maximum divergence. Trapezoid-model divergence conjecture. The divergence from the trapezoid model \Ta,b,d\T_{a,b,d} is at least (a+b1)log2(a+b-1)\log 2 and at most (a+bd+b1)log2(a+bd+b-1)\log 2. This upper bound is sharp if and only if d=0d=0. The preceding theorem gives a related upper bound, while the conjecture compares the trapezoid model with the two surrounding box models; the stated sharpness condition remains to be established.

Sources & referencesView supporting material

Primary source

Yulia Alexandr and Serkan Hoşten, “Maximum information divergence from linear and toric models”, arXiv:2308.15598 (2023).

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.