Abstract
The true slime mould, Physarum polycephalum, develops as a vascular network of protoplasm, connecting node-like sources of food in an effort to solve multi-objective transport problems. The organism first establishes a dense and continuous mesh, reinforcing optimal pathways over time through constructive feedbacks of protoplasmic streaming. Resolved vascular morphologies are the result of an evolutionarily-refined mechanism of computation, which can serve as a versatile biological model for network design at the urban scale. Existing digital Physarum models typically use positive reinforcement mechanisms to capture meshing and refinement behaviours simultaneously. While these automations generate accurate descriptions of sensory and constructive feedback, they limit stepwise design control, reducing flexibility and applicability. A model that decouples the two “phases” of Physarum behaviour would enable multistage control over network growth. Here we introduce such a system, first by producing a site-responsive mesh from a population of nutrient-attracted agents, and then by independently calculating from it a flexible, proximity-defined shortest-walk to produce a final network. We develop and map networks within existing urban environments that perform similarly to those biologically grown, establishing a versatile tool for bio-inspired urban network design.
Highlights
Slime mould can crucially serve as a biological model for adaptive computational network d esign[20]
In the 1960s, architects at the Institute for Lightweight Structures in Stuttgart began experimenting with self-assembling natural materials, hoping to find generalizable physically-optimized solutions for urban network design
We arranged a collection of oats on otherwise nutrient-poor agar plates and placed a small sample of the active plasmodium in the center of each plate
Summary
Slime mould can crucially serve as a biological model for adaptive computational network d esign[20]. While the organism can impressively reconstruct existing anthropically-designed urban n etworks[20], the very discrepancies between city infrastructures, and their biologically-grown Physarum analogues, have proven crucial for understanding characteristics of both urban network planning and Physarum intelligence[20,21,32,46]. As a growth point moved along a nutrient-poor terrain under the influence of a field source. In another model, two agent-like Physarum populations searched across a nutrient-populated domain, sampling for chemo-nutrients and existing trails (i.e., regions where the cell had already occupied)[50]. Networks started as randomly meshed lattices, sidestepping initial foraging behaviour, but still evolved to solve transportation problems
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
