Parallel Acceleration of Slime Mould Discrete Models

Biological organisms have become an inspiration for many computer scientists in order to process and analyze complex engineering problems. A well known example of this success story departs from the application of plasmodium of Physarum to solving the shortest path problem as experimentally demonstrated in the case of a labyrinth as well as to other graph related problems. There are many modeling tools trying to mimic the behavior of Physarum. We consider a discrete and parallel model, namely cellular automata (CA) based model implemented in hardware , which attempts to describe and, moreover, mimic the Physarum’s behavior in a maze . In order to take full advantage of the CA inherent parallelism, we implemented the model on a Field Programmable Gate Array (FPGA) . Two implementations were considered in order to accelerate the model’s response and improve the exactness of the experimental results. Their main difference subsists in the precision produced by the numerical representation of CA model parameters. The modeling efficiency of both approaches was compared depending on the resulting error propagation. The presented FPGA implementations accelerate considerably the performance of the CA algorithm when compared with its software based version. Finally, a Graphical Processing Unit (GPU) will exploit the prominent feature of parallelism that CA structures inherently possess in contrast to the serial computers, thus accelerating the response of the proposed model in a more easy to be programmed fashion. As a result, these implementations can also be considered as a preliminary, parallel and accelerated CA-based Physarum Polycephalum hardware virtual lab, which reproduces the characteristics of the biological organism towards its application to the shortest-path problem and thus increases significantly the computational speed.