[en] Biological membranes are examples of 'smart' materials whose properties and behaviour emerge from the propagation across many scales of the molecular characteristics of their constituents. Artificial smart materials, such as drug delivery vehicles and biosensors, often rely on modifying naturally occurring soft matter, such as polymers and lipid vesicles, so that they possess useful behaviour. However, the complexity of natural membranes, both in their static properties, exemplified in their phase behaviour, and in their dynamic properties, as in the kinetics of their formation and interactions, hinders their rational modification. Mesoscopic simulations, such as dissipative particle dynamics (DPD), allow in silico experiments to be easily and cheaply performed on complex, soft materials requiring as input only the molecular structure of the constituents at a coarse-grained level. They can therefore act as a guide to experimenters prior to performing costly assays. Additionally, mesoscopic simulations provide the only currently feasible window on the length- and timescales relevant to important biophysical processes such as vesicle fusion. We review here the development of computational models of bilayer membranes, and in particular the use of mesoscopic simulations to follow the molecular rearrangements that occur during membrane fusion