The steady-state distributions of coagulation-fragmentation processes (Q1127641): Difference between revisions

A continuous-time homogeneous Markov chain model is proposed for the coagulation-fragmentation (CF) process of \(N\) particles, which has applications in biology, physical chemistry and astrophysics. The model restricts simultaneous coagulations and fragmentations to two groups and assumes ''perfect missing''. The chain is ergodic under mild conditions. From the stationary distribution, the authors obtain easily the expected group size distribution. The case of truncation to groups of maximum size \( L<N\) is also considered. Exact results are only possible for small \(N\). For large \(N\), the author proposes a Monte Carlo simulation method based on the embedded discrete-time Markov chain, which gives very good approximations. Rather than to use the classic deterministic models of CF, the author uses the Markov chain model to derive deterministic discrete CF equations for the expected group sizes. In order for them to be useful, an approximation (obtained by cancelling the correlations between quadratic terms) is used. This approximate deterministic model and the classical deterministic models are used in some pair formation (truncation to size two groups) and animal grouping models, and are shown to perform badly for small \(N\) and reasonably well for some (but not all) cases of large size \(N\).