Uniqueness and iteration conjecture for the fixed point of U^,0\hat{U}^{\dagger,0}

Let U^,0:RNRN\hat{U}^{\dagger,0}:\mathbb{R}^N\to\mathbb{R}^N be the operator defined for the unknown stochastic shortest path problem, and let CPprimalUnknown,1\mathrm{CP}_{\mathrm{primalUnknown}}^{\dagger,\ell_1} denote the associated optimisation program. A fixed point of U^,0\hat{U}^{\dagger,0} is a vector xRNx\in\mathbb{R}^N satisfying U^,0x=x\hat{U}^{\dagger,0}x=x. Uniqueness and iteration conjecture. There exists a unique fixed point of U^,0\hat{U}^{\dagger,0}, which is equal to the unique optimal solution to CPprimalUnknown,1\mathrm{CP}_{\mathrm{primalUnknown}}^{\dagger,\ell_1}. In almost all cases, this can be found by iterating the operator U^,0\hat{U}^{\dagger,0} starting at any point in RN\mathbb{R}^N. This is presented as an open question; the source does not establish the claimed uniqueness or convergence, and notes that additional assumptions such as all policies being proper may be needed.

Sources & referencesView supporting material

Primary source

Kelli Francis-Staite, “Convex duality for stochastic shortest path problems in known and unknown environments”, arXiv:2208.00330 (2022).

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.