RT Journal Article
SR Electronic
T1 Vectorization techniques for efficient agent-based model simulations of tumor growth
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 032086
DO 10.1101/032086
A1 Jan Poleszczuk
A1 Heiko Enderling
YR 2015
UL http://biorxiv.org/content/early/2015/11/17/032086.abstract
AB Multi-scale agent-based models are increasingly used to simulate tumor growth dynamics. Simulating such complex systems is often a great challenge despite large computational power of modern computers and, thus, implementation techniques are becoming as important as the models themselves. Here we show, using a simple agent-based model of tumor growth, how the computational time required for simulation can be decreased by using vectorization techniques. In numerical examples we observed up to 30-fold increases in computation performance when standard approaches were, at least in part, replaced with vectorized routines in MATLAB.