A Survey on Physarum Polycephalum Intelligent Foraging Behaviour and Bio-Inspired Applications

AbuBakr Awad, Wei Pang, David Lusseau, George M. Coghill

Introduction

For decades, Physarum polycephalum—a single-celled slime mould—has captivated computer scientists with its ability to solve complex spatial problems without a central nervous system. From navigating mazes to efficiently reconstructing the Tokyo railway network, this organism serves as a biological computer. While traditional research focuses on solitary agents, this survey explores the transition toward multi-agent competition and decentralized network optimization, offering new strategies for solving NP-hard problems and designing robust infrastructure.

Biological Foundations and Decentralized Intelligence

Physarum polycephalum processes information through the synchronous oscillations of its cytoplasm. It lacks a brain, yet it compares multiple options and integrates "memory" through extracellular slime trails. Its efficiency is driven by a simple positive feedback loop: protoplasmic tubes that transport more nutrients grow thicker, while those with low flow eventually wither. This natural mechanism allows the organism to solve Steiner tree and shortest-path problems with remarkable precision.

The life cycle of Physarum polycephalum and its plasmodial network in a petri dish.

Modeling Approaches

The survey categorizes several modeling frameworks used to simulate these biological processes:

  • Flow-conductivity models: Based on Hagen-Poiseuille and Kirchhoff Laws to simulate vein thickening.
  • Reaction-diffusion (RD) models: Treating the organism as a biological chemical computer.
  • Cellular Automata (CA): Utilizing discrete grid-based rules to simulate growth.

The authors highlight a novel Hexagonal Cellular Automata and RD system. Unlike standard square grids, hexagonal grids better approximate natural circular diffusion and incorporate biological mechanisms like allorecognition.

Multi-Agent Competition and Allorecognition

A critical frontier in slime mould research is "competition modelling." When multiple Physarum agents interact, they exhibit allorecognition—the ability to distinguish "self" from "other" via chemical signals in their slime sheaths. This mechanism allows different strains to avoid each other rather than fusing, facilitating decentralized network partitioning.

Conceptual model of non-contact allorecognition between two plasmodia via slime sheath signals.

Competition acts as a global optimizer; multiple agents exploring an environment prevent the system from getting trapped in a "local optimum" (a single, potentially sub-optimal solution) too early. This multi-agent perspective is essential for modeling real-world challenges like mobile wireless sensor networks or complex supply chains.

Applications and Benchmarks

Physarum-inspired algorithms are competitive with traditional computational methods, such as Dijkstra’s algorithm, in specific contexts:

  • Infrastructure Design: Models have successfully approximated major highway systems in Mexico, Iberia, the USA, and Germany with high tolerance for dynamic changes.
  • Hybrid Algorithms: Integrating Physarum logic into Ant Colony Optimization (ACO) or Genetic Algorithms (GA) improves efficiency and prevents premature convergence.
  • Biological Computing: The development of "Physarum chips" uses the organism as a living transducer for logic gates and arithmetical computing.

Conclusion and Future Outlook

While the biological growth of Physarum is slow compared to silicon-based processing, its principles offer a blueprint for the next generation of decentralized software. Future research is directed toward Physarum-inspired sensors, where the organism's biological response to toxins or environmental changes is converted into electrical signals. Refining multi-agent competition models could revolutionize the management of decentralized robotic swarms or sensor networks in unpredictable environments.

Summary updated

BibTeX
@article{awad2022survey,
  author = {AbuBakr Awad and Wei Pang and David Lusseau and George M. Coghill},
  title = {{A Survey on Physarum Polycephalum Intelligent Foraging Behaviour and Bio-Inspired Applications}},
  journal = {Artificial Intelligence Review},
  year = {2022},
  url = {https://openreview.net/forum?id=F11T9m2vBk},
}
Figure