The TTC subway system is notoriously inefficient — but solving that is a no-brainer for the brainless slime mould P. polycephalum

This may seem unlikely — what do these primitive, blob-like creatures know about urban design? It turns out that the types of networks that P. polycephalum have developed naturally over the course of their evolutionary history are surprisingly similar to the kinds of networks that humans are interested in replicating. 

Raphael Kay, a graduate student in materials science and engineering at U of T, was one of the lead authors of a study recently published in Scientific Reports that investigated this connection in more detail. In a conversation with The Varsity, Kay talked about the study’s discoveries and how to translate virtual slime mould models into practical public transit applications.

The complexity of a single-celled blob

These blob-shaped protists have the marvellous ability to form beautifully intricate networks without any kind of recognizable brains — which makes them a promising inspiration for computational models. 

To maintain the integrity of their body’s fluid network, each of these organisms are required to make a series of decisions about how to navigate their environment. Scientists are interested in investigating the underlying mechanisms that equip them with the ability to perform multi-step computations, even in the absence of a refined neural circuit. 

What is significant about these blobs is the particular way that they attempt to optimize their use of exhaustible resources such as food. Humans are not as good at optimizing resources as these simple-minded creatures — but the blobs’ biological approach to resource management does show similarities to ways that we design our cities and roadways. 

In an attempt to improve the future of urban networks, Kay and his colleagues at U of T’s Functional and Adaptive Surfaces Group created digital simulation models to replicate the protist’s optimization processes.

All about the right balance 

Kay and his colleagues chose three parameters to optimize network efficiency: cost, travel time, and vulnerability. The behaviour of slime moulds offers insights into how researchers can maintain a balance between these network parameters. This is a difficult problem: for example, a transport system design that focuses primarily on lessening travel time may end up becoming an overly sophisticated network that is incomprehensible to passengers. 

But how do blob simulations improve network design about the TTC’s costs and travel times? 

These results are actually an example of the power of the virtual blob model at its finest. The answers to network optimization questions rely on the concepts of ‘resiliency’ and ‘vulnerability.’ In simple terms, Kay explained, the vulnerability of a network refers to how much longer it would take to move from one station to another if one link were removed from the network as a whole. In this scenario, resiliency is simply the inverse of vulnerability — a network’s resiliency refers to its ability to recover from the removal of one of its links.

The networks generated using the slime mould model are projected to be about 40 per cent less vulnerable than the current real-life networks designed for similar transportation problems. This result is only possible because of the dynamics between resilience and vulnerability. 

From blob to urban network blueprint 

Kay explained that their results were a practical translation of what the slime moulds came up with. Kay drew inspiration from the blob’s interest in connecting food sources and translated this idea to connecting people in an urban transportation network — although unlike the blob protists, commuters are more concerned about efficiency of the TTC subway system as a whole than about finding the most effortless way to navigate from an eatery to the nearest café. 

The most updated version of the virtual slime mould model is developed based on how the blobs act in response to “attractiveness” of food sources. The object of attraction, in the blob’s case, is always food. Not only do the blobs navigate toward their intended food sources, but they also avoid unwanted harmful substances by circumventing contaminated parts of their petri dishes. The researchers created a visual manifestation based on this kind of movement, Kay added, which offered them a unique perspective on “[making] a more dynamic knob for tuning how the model behaves [in real-world situations].”

Kay recalled the research team’s first attempt to transfer the computational slime mould model into practice. The team compared their virtual slime model to the arrangement of sidewalks in Canada’s Wonderland to figure out the most efficient walkways to go from ride to ride, with seemingly successful results. 

Their experiment in the amusement park led Kay to compare his slime mould model to the TTC subway. This time around, the researchers overlaid the ‘ideal’ network they generated via the simulation model on top of the actual TTC map for spatial analysis, which showed quantitative differences between the two overlapping network designs. 

This line of work produced objective quantitative data that could serve as a guideline for future retrofitting projects of the TTC. Kay’s effort to implement his research in tangible and relevant contexts is a fine example of real-world research applications, and can serve as encouragement to students who are looking to contribute to the world beyond academia. 

Final reflections

More often than not, people believe that models inspired by biology must be the gold standard, and that they’re suitable to resolve issues for a wide variety of situations. In reality, though, the urban design ideas generated by this virtual model can only serve as a reference for engineers working to optimize the efficiency of public transport and similarly structured networks. 

When asked about plans for future studies, Kay expressed wishes for his model to be able to assist in improving urban infrastructure in highly urbanized areas, or to potentially design a first pass for engineers who are developing a transport network from scratch. Kay is excited to use his model to evaluate the weaknesses of existing systems and generate tangible suggestions for  retrofitting urban networks. 

In his undergraduate studies in John H. Daniels Faculty of Architecture, Landscape and Design, Kay appreciated the opportunity to learn from world-renowned architects who were much older than him. In Kay’s opinion, this directly connects to his current research, as well. Thinking about the experience and knowledge he got from older human architects made him wonder what would happen if he pushed the idea further — “Why not go all out to ask [these ancient] architects to design a city?” 

Editor’s note (27 March): A previous version of this article incorrectly stated that the paper was published in the journal Nature. In fact, it was published in the journal Scientific Reports.