Cellular Automata Models Simulating Slime Mould Computing

Slime mould computers have been used to solve graph-theoretical problems like mazes and evaluate man-made transport networks. For the laboratory experiments that demonstrate these computing capabilities, slime mould is first starved and then introduced to an area with attractants placed on key positions. The behaviour of slime mould during these laboratory experiments have been simulated by a model based on cellular automata (CAs). The advantages of a software model over the real slime mould are repeatability and faster productions of results. Using CAs can be justified by the emergence of global behaviour from local interactions, a rule that applies also on the real slime mould. The results of the model have been compared to the ones produced during laboratory experiments and found in good agreement both for maze solving and network designing. After thorough examination of the laboratory experiments an updated model was developed, which yielded more efficient networks. As the model was parametrized to produce slightly differentiated results, the effects of these parameters were studied.