From Pattern Formation to Material Computation: Multi-agent Modelling of Physarum Polycephalum
Author: Jeff Jones
Springer. Emergence, Complexity and Computation Series. May, 2015
Reviewed by Andrew Adamatzky
A recent addition to Springer’s popular “Emergence, Complexity and Computation” series, From Pattern Formation to Material Computation by Jeff Jones is a monograph which explores the author’s research into modelling the behaviour of the true slime mould Physarum polycephalum, an apparently primitive single-celled organism which has captured the imagination of biologists for over one hundred years. Since the start of this century it has also been thrust into the computational spotlight as one of the simplest living examples of complex distributed computing. From initial reports of maze solving the organism has been applied to shortest path problems, proximity graphs, plane division problems, combinatorial optimisation, and even a primitive memory (for an up-to-date compendium of recent research the interested reader can refer to the forthcoming ). Its computational feats are all the more remarkable because the organism is composed of such simple component parts (a multi-nucleate membrane bound syncytium, in the plasmodial stage of its life-cycle). It has been described as a living, membrane-bound reaction-diffusion processor , equivalent to — and potentially surpassing — the performance of prototype chemical processors developed in recent decades .
The book attempts to reproduce the complex behaviour of P. polycephalum by using the same limitations faced by the organism itself: simple component parts, coupled with local interactions to result in complex emergent behaviours. Such simple local interactions are, of course, a popular recipe to explore emergent behaviours in other simple systems, as cellular automata. The book begins with an overview of the organism and a description of previous biological and computational research featuring the plasmodium. After giving an overview of classical and unconventional computation the reasoning behind the modelling approach is described.