Procedure conjecture for the fixed point of L^π,0\hat{L}_{\pi}^{\dagger,0}

From papers

Let P0,P1,P2,P11,P12,P21,P22P_0,P_1,P_2,P_{11},P_{12},P_{21},P_{22} be the transition matrices defining the active pieces of L^π,0\hat{L}_{\pi}^{\dagger,0}, let x0,x1,x2,x11,x12,x21,x22x_0,x_1,x_2,x_{11},x_{12},x_{21},x_{22} be their respective fixed points, and let JJ^* be the fixed point of L^π\hat{L}_{\pi}. The feasible region is determined by x0xiJx_0\leq x_i\leq J^*, together with membership of xix_i in the corresponding active piece. Procedure conjecture. The fixed point of L^π,0\hat{L}_{\pi}^{\dagger,0} can be found by calculating these fixed points, discarding any xix_i that fails either feasibility test, and then choosing whichever of P1P_1 or P2P_2 remains; if neither remains, choose the remaining fixed point whose elements have the highest sum. The source gives this as a proposed procedure based on the preceding two-state analysis, without establishing it as a theorem.

Progress summary

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

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).

Solutions 0

No solutions have been posted yet.