Growing spanning trees in plasmodium machines

Purpose - The purpose of this paper is to address the novel issues of executing graph optimization tasks on distributed simple growing biological systems. Design/methodology/approach - The author utilizes biological and physical processes to implement non-classical, and in principle more powerful, computing devices. The author experimentally verifies his previously discovered techniques on approximating spanning trees during single cell ontogeny. Plasmodium, a vegetative stage of slime mold Physarum polycephalum, is used as experimental computing substrate to approximate spanning trees. Points of given data set are represented by positions of nutrient sources, then a plasmodium is placed on one of the data points. Plasmodium develops and span all sources of nutrients, connecting them by protoplasmic strands. The protoplasmic strands represent edges of the computed spanning tree. Findings - Offers experimental implementation of plasmodium devices for approximation of spanning tree. Practical implications - The techniques, discussed in the paper, can be used in design and development of soft bodied robotic devices, including gel-based robots, reconfigurable massively robots, and hybrid wet-hardware robots. Originality/value - Discusses original ideas on growing spanning trees, and provide innovative experimental implementation.