1: \begin{abstract}

2: We study a class of Markovian systems of $N$ elements taking values in $[0,1]$ that evolve

3: in discrete time $t$

4: via randomized replacement rules based on the ranks of the elements. These rank-driven processes

5:  are inspired

6: by variants of the Bak--Sneppen model of evolution, in which the system represents an evolutionary

7: `fitness landscape' and which is famous as a simple

8: model displaying self-organized criticality.

9: Our main results are concerned with long-time

10: large-$N$ asymptotics for

11:  the general model in which, at each time step,

12: $K$ randomly chosen elements are discarded and replaced by independent

13: $U[0,1]$ variables, where the ranks of the elements to be replaced are chosen, independently at each time step,

14: according to a  distribution $\kappa_N$ on $\{1,2,\ldots,N\}^K$. 

15: Our main results are that, under   appropriate conditions on $\kappa_N$, 

16: the system exhibits threshold behaviour at $s^* \in [0,1]$, where $s^*$ is a

17: function of $\kappa_N$, and

18: the marginal distribution of a randomly selected element converges to $U[s^*, 1]$ as

19: $t \to \infty$ and $N \to \infty$. 

20: Of this class of models, results in the literature have previously been given for 

21: special cases only, namely the `mean-field' or `random neighbour' Bak--Sneppen model. Our proofs 

22: avoid the heuristic arguments of some of the previous work and   use   Foster--Lyapunov ideas. Our results extend existing results

23: and establish their natural, more general context. We derive some more specialized results for the particular case where $K=2$.

24: One of our  technical tools is a result on convergence of stationary distributions

25: for families of uniformly ergodic Markov chains on increasing state-spaces, which may be of independent interest.

26: \end{abstract}