Physarum Network Simulation

Brief Description:

Physarum Polycephalum, also known as slime mold is a single celled organism which displays a kind of swarm intelligence through it's ability to form efficient food networks. The mold does not have a central control mechanism but instead relies on local interactions within its cellular structure to efficiently develop its network. This allows the slime mold to adapt to its environment without a central decision making process.

This simulation was created to replicate how interactions between individual agents within a network can result in complex emergent behaviours. The simulation uses pheromone trails to mimic the protoplasm deposited during the formation of a slime mold network. The map is initially populated with 6000 agents in random orientation, as each agent moves it leaves behind it a pheremone trail that diffuses and evaporates at a constant rate. When an agent encounters a previously laid trail it interacts with it via three sensors under a set of predetermined rules. This behaviour demonstrates how the slime mold is able to reinforce successful paths that lead to nutrients using local interactions. By altering the agent and environment parameters listed below you can observe the network transforming as each individual agent interacts with the trails in different ways. Use the buttons above or the keybinds listed below to alter the parameters.

To adjust agent parameters:

• A/S = adjust sensory angle
• O/P = adjust sensory offset
• T/Y = adjust turn angle

To adjust environment parameters:

• D/F = adjust diffusion
• E/R = adjust evaporation
• LMB/RMB/wheel: erase/deposit food/pheromone trail
• ↑/↓: decrease/increase timestep

Parameter Notes

• The sensory angle is the breadth of each agent's sensory perception, especially in relation to directional sensing. A wider sensory angle allows the agent to detect stimuli over a broader range.
• Sensory offset is the distance ahead of each agent where it is able to sense it's environment.
• Higher trail diffusion leads to wider, more dispersed trails. This affects the likelihood of agents encountering and following these trails.
• Higher trail evaporation rates lead to shorter lived trails which in turn requires agents to constantly explore new paths and avoid relying on outdated trails.