Code for my research projects are available online through github and figshare. Below are the main Python packages I have been working on developing over the last few years.
Tensor Networks for Stochastic Models
Harvesting the power of tensor networks for representing high-dimensional probability distributions of stochastic systems. StochasticTN is a python library for constructing Matrix Product State representations of probability distributions arising from Markov processes such as the contact process (or SIS model for epidemiologists).
In this package, the tensor network directly represents a non-negative probability distribution and not a quantum mechanical wavefunction. DMRG methods are implemented to find the leading eigenvectors of the generator of the Markov process, which represents the late-time steady state distribution. The generator may be tilted to find the scaled cumulant generating function of the dynamical activity, used to study the large deviation statistics in the model. More details are described in this publication.
Exact solutions of the SI model on networks
The logistic differential equation of Verhulst is a landmark result in many areas of science concerning growth, such as population growth or information spreading processes. We analyze the exact Markovian system describing logistic growth on a network, otherwise known as the SI model. Exact solutions can be organized in terms of contributions from subgraphs of the network of interest and we present here a general code for computing these contributions systematically and analytically, based on SymPy and NetworkX. More details are described in this publication.
Wout Merbis 