An Adaptive Agent-Based Model of Homing Pigeons: A Genetic Algorithm Approach

Abstract

Conventionally, agent-based modelling approaches start from a conceptual model capturing the theoretical understanding of the systems of interest. Simulation outcomes are then used "at the end" to validate the conceptual understanding. In today's data rich era, there are suggestions that models should be data-driven. Data-driven workflows are common in mathematical models. However, their application to agent-based models is still in its infancy. Integration of real-time sensor data into modelling workflows opens up the possibility of comparing simulations against real data during the model run. Calibration and validation procedures thus become automated processes that are iteratively executed during the simulation. We hypothesize that incorporation of real-time sensor data into agent-based models improves the predictive ability of such models. In particular, that such integration results in increasingly well calibrated model parameters and rule sets. In this contribution, we explore this question by implementing a flocking model that evolves in real-time. Specifically, we use genetic algorithms approach to simulate representative parameters to describe flight routes of homing pigeons. The navigation parameters of pigeons are simulated and dynamically evaluated against emulated GPS sensor data streams and optimised based on the fitness of candidate parameters. As a result, the model was able to accurately simulate the relative-turn angles and step-distance of homing pigeons. Further, the optimised parameters could replicate loops, which are common patterns in flight tracks of homing pigeons. Finally, the use of genetic algorithms in this study allowed for a simultaneous data-driven optimization and sensitivity analysis.

ABWiSE v1.0: toward an agent-based approach to simulating wildfire spread

“…We base the flammability and fuel of cells on the Canadian Forest Fire Danger Rating System (CFFDRS) (Van Wagner, 1974;Wotton, 2009). We chose it due to the availability of fuel type data in Canada and the system's use around the world (Opperman et al, 2006). The system includes 16 classes of vegetation for which there are empirically derived equations relating fuel moisture and weather to fire behaviour.…”

“…Fire is an integral part of ecosystems the world over but also poses a serious danger to human life and property (Bowman et al, 2011;Moritz et al, 2010;Brenkert-Smith et al, 2013;Butry et al, 2001;Carroll et al, 2006;Chuvieco et al, 2014;Kochi et al, 2010;Richardson et al, 2012). In recent years, anthropogenic climate change has exacerbated this danger chiefly by lengthening growing seasons and increasing the risk of drought (Flannigan et al, 2016;Lozano et al, 2017), leading to more frequent and more extreme fires in many parts of the world (Chuvieco et al, 2016;Kirchmeier-Young et al, 2019, 2017. The use of controlled burning has, for a very long time (Gott, 2005;Roos et al, 2021), helped to mitigate the risks of extreme fires and to maintain forest health (Boer et al, 2009;Camp and Krawchuk, 2017;Fernandes and Botelho, 2003).…”

From Forecasting to Real-Time Process Controlling with Continuous Simulation Model Updates

“…However, the integration of real-time data into simulation models enables the possibility of comparing simulations to the real system during the project's execution. The incorporation of real-time data into agent-based models improves the predictive ability of such models and results in increasingly wellcalibrated model parameters and more accurate outcomes (Oloo & Wallentin, 2017).…”

Transferring decision boundaries onto a geographic space: Agent rules extracted from movement data using classification trees

“…Classification trees are recursive algorithms ideally suited to explore data structures as well as analyse complex ecological data (Loh, 2011). Interesting work using such technologies to identify potential environmental factors underlying animal movement, determine population distributions, or predict zoonotic disease transmission risk is already well documented (Ahearn, Dodge, Simcharoen, Xavier, & Smith, 2017; Elith, Leathwick, & Hastie, 2008; Han, Schmidt, Bowden, & Drake, 2015; Leathwick, Elith, Francis, Hastie, & Taylor, 2006; Oloo & Wallentin, 2017; Torrens, Li, & Griffin, 2011; Ward, Evans, & Malleson, 2016). With respect to using these technologies to inform agent‐based models, notable examples include: context‐sensitive random walks that incorporate local external factors to simulate the movement of tigers at Royal Chitwan National Park in Nepal (Ahearn et al., 2001); genetic algorithms to simulate representative relative‐turn angles and step‐distance of homing pigeons (Oloo & Wallentin, 2017); reinforcement learning to contextualize the risk and reward of agent behaviour (Sutton & Barto, 1999; Tang & Bennett, 2010); and artificial neural networks to assign weights to link environmental features to an agent's internal spatially explicit map of its surroundings (Huse, Strand, & Giske, 1999; Strand, Huse, & Giske, 2002).…”