Adaptive Path-Finding and Transport Network Formation by the Amoeba-Like Organism Physarum

Abstract

Abstract. The giant amoeba-like plasmodia of Physarum is able to solve the shortest path through a maze and construct near optimal functional networks between multiple, spatially distributed food-sources. These phenomena are interesting as they provide clues to potential biological computational algorithms that operate in a de-centralized, singlecelled system. We report here some factors that can affect path-finding through networks. These findings help us to understand more generally how the organism tries to establish an optimal set of paths in more complex environments and how this behaviour can be captured in relatively simple algorithms.

Incremental transformation of spatial intelligence from smart systems to sensorial infrastructures

“…New applications/methods of computation include photonic computation (Erişen, 2018, p. 144;Tait et al, 2014) for investigating the entanglement of green living, and bacteria, among other topics, as well as molecular computation for tracking the motion of particles (Kunita et al, 2013;Suzuki & Nakagaki, 2013). Endeavours for revealing the action of particles in built environments (Bhattacharya et al, 2020), for instance, provide substantial evidence or even ground truth for air quality to progress further by urban computing (Zheng et al, 2014).…”

A growth model for water distribution networks with loops

“…Physarum polycephalum has shown various intelligent behaviour such as finding the shortest paths [26], adapting to changing environments [27] and building high-quality networks that realize a feasible balance between cost, efficiency and robustness [24]. Previous studies suggested that Physarum polycephalum may inform the design of next-generation, adaptive and robust spatial infrastructure networks with decentralized control systems [28][29][30].…”

A Model for Simulating Emergent Patterns of Cities and Roads on Real-world Landscapes

“…The resultant Landscape-Transport-Population (LTP) reinforcement model has three key elements: (i) Detailed landscape topography at 90 m resolution; (ii) The effective transport distance across the real-world landscape; and (iii) Turing-like pattern formation driven by flow-based reinforcement for population distributions. Our approach draws on our previous works [23][24][25][26][27] where a biologically-inspired, current-reinforcement rule can construct realistic transport networks between food-supply points under physical and physiological constraints, and the principle of co-evolution of nodes and links that dynamically organize scale-free networks via a diffusion process 28,29 .…”