Abstract

Agent-based modelling has been shown to be a valuable method for modelling systems whose behaviour is driven by the interactions of distinct entities. However, the methodology faces a fundamental difficulty: there are no established mechanisms for dynamically incorporating real-time data into models. This work begins to address this gap by demonstrating how data assimilation techniques can be used to allow data to stream into running models to reduce the uncertainty in their estimates of the current state of the target system. This allows for better predictions of the present and potentially more accurate short-term forecasts. The methods are illustrated using a case study of pedestrian movements. By laying the groundwork for the real-time simulation of crowd movements, this work has implications for the management of complex environments such as transportation hubs, hospitals, shopping centres, etc.

Publications

Kieu, Le-Minh, Nicolas Malleson, and Alison Heppenstall. 2020. Dealing with Uncertainty in Agent-Based Models for Short-Term Predictions. Royal Society Open Science 7 (1): 191074. DOI: 10.1098/rsos.191074.

Malleson, Nick, Kevin Minors, Le-Minh Kieu, Jonathan A. Ward, Andrew A. West, and Alison Heppenstall. 2019. Simulating Crowds in Real Time with Agent-Based Modelling and a Particle Filter. ArXiv:1909.09397 [Cs], September. http://arxiv.org/abs/1909.09397.

Ward, Jonathan A., Andrew J. Evans, and Nicolas S. Malleson. 2016. Dynamic Calibration of Agent-Based Models Using Data Assimilation. Royal Society Open Science 3 (4). DOI: 10.1098/rsos.150703.

But that's too hard, so we're currently looking at:

Real Time Crowd Modelling

Context: simulate a crowd in real time

What methods can we use to incorporate data?

How much data do we need?

Track every individual?

Track some individuals?

Just aggregate counts (e.g. number of people passing a footfall camera)

Experiments with a simple, hypothetical train station (StationSim)

Uncertainty in StationSim

Only one source of uncertainty: collision avoidance

If a fast agent reaches a slow agent, they try pass to the left or right

Next slides: 2 examples of data assimilation on this simple model

Example 1: Particle Filter

A 'brute force' ensemble method

No Gaussian assumptions

Create N realisations of the model ('particles')

Run all the particles forward in time until you receive some new data

Compare the particles to the observation and:

Weight each particle depending on how close it is to the observations

Re-sample the population of particles using the weights (good particles are kept, bad ones disappear)

Repeat

Experimental Setup

'Identical twin' experiment: known 'real world' conditions

Crowd Simulation with a Particle Filter

Crowd Simulation with a Particle Filter

Particle Filter Results

Particle Filter Results

Example 2: Unscented Kalman Filter

Similar to the (very popular) Ensemble Kalman filter

Should be more efficient

But assumes Gaussian distributions

A few sigma points are chosen to represent the model state

Then some complicated maths happens ...

UKF Results

Observe a proportion of the agents

UKF Results

Observed v.s. Unobserved agents

Example 3:
Data Assimilation on a Bus Simulation

Context: simulate bus routes in real time

We have GPS bus positions, but to make good term forecasts we need to be able to infer other factors

Number of people waiting at bus stops

Number of people on the bus

Surrounding traffic levels

Etc.

Aim: test a particle filter as the means of assimilating real-time GPS positions into a model.

Bus Simulation

Bus Simulation with a Particle Filter

Ethical Implications

Data Bias

Need to be very careful: biased data -> biased models

The digital divide

Tracking People

Advantage with these methods is we don't need to track people

Models work with counts or aggregate flows

Conclusions

(Agent-Based) simulations of urban systems will benefit from being able to react to up-to-date data

Otherwise uncertainty will cause them to diverge from reality

This work has begin to experiment with some mathematical methods that might be able to conduct data assimilation on agent-based models.

Ultimately work towards a digital twin of urban flows