I have a discrete time stochastic process where an interval from 0 to L consists of smaller sub-segments, where the boundaries are always at integers, and L is some large integer. At each iteration, a segment is chosen with probability proportional to its length, and an operation is performed on it, such as growing it while shrinking the neighboring segment, shrinking it while growing the neighboring segment, or cutting it into two smaller segments. The quantitative details are not important here.
What analytic approach can I use for creating probabilities of the positions of boundaries over time? It seems difficult since the number of boundaries can change. I am familiar with Markov chains, but I'm not sure what the state space would be. I can certainly model this numerically, but I'm looking for a theoretical approach.
Even if you don't know that answer, saying "this is similar to a XX process" would be great.