Agent-Based High-Performance Simulation of Biological Systems on the GPU

Simulation of biological systems are computationally demanding due to the large scale reaction networks of bacterial cells. This scalability issue escalates, in particular, when bacterial colonies, formed by many individual cells, are simulated. Agent-based modelling environments on parallel architectures, such as the FLAME (Flexible Large-scale Modelling Environment) framework, are good candidates to simulate such systems, but due to the complex nature of cellular systems more advance technology is needed. In this paper, we utilise FLAME GPU, extending FLAME with a high performance graphics processing unit, to simulate a pulse generator, a typical multicellular synthetic biology system. This system is specified using a membrane computing model. We also illustrate the performance improvement of FLAME GPU over FLAME.