Estimation of the error-components model with incomplete panels (Q5966490)

To vertex i of an undirected graph G there is assigned an integer \(x_ i(t)\); \(J_ i\) is the set of vertices adjacent to i, \(d_ i=| J_ i|\), \((i=1,2,...,n\); \(t\geq 0)\). Then \(x_ i(t+1)\) is defined to be \(x_ i(t)-1\) if \(\sum_{j\in J_ t}x_ j(t)<d_ ix_ i(t)\), \(x_ i(t)\) if \(x_ j(t)=x_ i(t)\) for all \(j\in J_ i\), and \(x_ i(t)+1\) otherwise. Theorem: For any initial assignment \(x_ 1(0),...,x_ n(0)\) there is a \(t_ 0\) such that for all \(t>t_ 0\) and all i, \(x_ i(t+2)=x_ i(t)\). ``This proves a generalization of a conjecture made by \textit{Ghiglia} and \textit{Mastin} [A cellular automata method for phase unwrapping, J. Optical Soc. (to appear)]. ``The proof we... give... is based on a modification of the proof used by \textit{E. Goles-Chacc}, \textit{F. Fogelman-Soulie}, and \textit{P. Pellegrin} [Discrete Appl. Math. 12, 261-277 (1985; Zbl 0585.94018)]... Furthermore... E. Goles (has) found a way to encode the iteration studied here into an iteration similar to those of'' (the last cited paper above).