Operadic approach to Markov processes with boundaries on the square lattice in dimension 2 and larger

D. Simon,

submitted ,



Markov processes are omnipresent in any dimension, especially in statistical mechanics; however their algebraic description is complete only in dimension 1 (which is often interpreted as time), for which linear algebra provides many tools complementary to the probabilistic approach: in particular invariant measures are eigenvectors. In larger dimension, such algebraic tools are absent, in particular because boundaries are more involved. We introduce new operadic tools to describe Markov processes in any dimension on the square lattice. The main consequence is a natural description of probabilistic boundary weights in terms of matrix product representations, for which we generalize the eigenvector property of 1D invariant measures. This new approach leads to new very interesting algebraic objects, with a local-to-global perspective. Moreover, finite size Markov processes on the square lattice can also be seen as a toy case of statistical Euclidean field theory: our algebraic tools are related to other categorical approaches to field theory, with the advantage of providing concrete computational tools.