Last passage percolation and traveling fronts (Q369673): Difference between revisions

The authors consider a stochastic process that was introduced by Brunet and Derrida to compute corrections for large but finite system size to some continuous limit equations in wave front propagation. It consists of a fixed number \(N\geq1\) of particles on the real line, initially at the positions \(X_1(0),\dotsc, X_N(0)\). The positions evolve according to the rule \[ X_i(t+1):=\max_{1\leq j\leq N}\{X_j(t)+\xi_{i,j}(t+1)\} \] (\(\{\xi_{i,j}(s), 1\leq i,j\leq N, s>0\}\) an i.i.d. family of real-valued random variables) giving an \(N\)-vector the components of which are not ordered. This vector describes the location of a population under reproduction, mutation and selection after \(t\) steps keeping the size constant. In relation to the current positions, the next ones are an \(N\)-sample of the maximum of the full set of the previous ones evolved by an independent step. It can be also viewed as a long-range directed polymer in a random medium with \(N\) sites. In this paper, the authors present some interesting results starting with Section 2 which contains some standard facts about the model. In Section 3, the wave front location in the case of the Gumbel law is given. In Section 4, they study the asymptotic behaviour as \(N\rightarrow\infty\) of the front profile (for the Gumbel law and small perturbations), and their relations to traveling waves and reaction-diffusion equations. In Sections 5 and 6, they expand the velocity in the case of integer-valued variables \(\xi_{i,j}\) (see above) bounded from above, starting with the Bernoulli case.