Different aspects of a random fragmentation model (Q2490065)

Fragmentation processes describe the evolution of an object that falls apart randomly as time passes. A random fragmentation model is formally introduced in the following way: Let \(S\) be the state space of decreasing numerical sequences which tend to 0, endowed with the uniform distance. Let \(X(t)\) be a Markovian process continuous in \(t\) with elements in \(S\). For \(r\geq 0\) let \(P_r\) denote the law started from the configuration \((r,0,\dots)\) and assume that the sizes of the pieces resulting at time \(t\) from an object with initial size 1 cannot exceed 1 and an object with 0 size cannot produce pieces with positive sizes. \(X\) is called a self-similar fragmentation if two conditions are fulfilled: (i) a real number \(\alpha\) exists such that for every \(r>0\) the distribution under \(P_1\) of the rescaled process \((rX(r^\alpha t)\), \(t\geq 0)\) is \(P_r\), and (ii) for every \(s\) in \(S\) and a sequence of processes \(X^{(i)}\) with law \(P_{si}\), rearranging the processes in decreasing order yields a version of the original process started from configuration \(s\). A homogeneous fragmentation is defined by \(\alpha=0\). The connections of fragmentation processes with (general) branching processes are discussed and several aspects of fragmentation processes are reviewed.