Agent-based modelling is becoming a widely used approach for simulating complex phenomena. By making use of emergent behaviour, agent based models can simulate systems right down to the most minute interactions that affect a system’s behaviour. In order to capture the level of detail desired by users, many agent based models now contain hundreds of thousands and even millions of interacting agents. The scale of these models makes them computationally expensive to operate in terms of memory and CPU time, limiting their practicality and use. This chapter details the techniques for applying Dynamic Hierarchical Agent Compression to agent based modelling systems, with the aim of reducing the amount of memory and number of CPU cycles required to manage a set of agents within a model. The scheme outlined extracts the state data stored within a model’s agents and takes advantage of redundancy in this data to reduce the memory required to represent this information. The techniques show how a hierarchical data structure can be used to achieve compression of this data and the techniques for implementing this type of structure within an existing modelling system. The chapter includes a case study that outlines the practical considerations related to the application of this scheme to Australia’s National Model for Emerging Livestock Disease Threats that is currently being developed.
TopIntroduction
In recent years, Agent-basedmodelling (ABM) has emerged as a robust technique for modelling complex, real-world phenomena. ABMs have been developed to simulate a broad range of natural and man-made systems including pedestrian movement, traffic movement, agricultural land use, social interactions, human and animal disease threats, and flocking behaviour in animals just to name a few (Berger, 2001; Busing & Mailly, 2004; D'Souza, Marino, & Kirschner, 2009; Elliston & Beare, 2006; Funk, Gerber, Lind, & Schillo, 1998; Miron, Garner, Donald, & Dyall, 2009; Perumalla & Aaby, 2008; Schelhorn, O'Sullivan, Haklay, & Thurstain-Goodwin, 1999; Strippgen & Nagel, 2009). Agent-based modelling has its origin in artificial intelligence applications where intelligent systems are designed around intelligentagents which are elements that perceive their environment through sensors and act upon their environment through actuators. The agent’s behaviour is defined by an agent function that takes its input from the agent’s sensors and calculates an action to be carried out using the actuators. This concept has been applied to the application of modelling by replicating entire complex phenomena as a system of multiple interacting agents. The agent-based modelling approach allows scientists to develop rich simulations capable of supporting experimentation at different conceptual levels within a complex system. This capability has led to the use of ABMs in a decision-support role for governments and industry, and fostered a demand for models with higher levels of detail within the individual agents and their interactions that provide an even greater scope for experimentation. In addition to this, ABMs are being applied to larger scale systems, modelling increasingly more complex phenomena.
An excellent example of this trend can be seen in Australia where agent-based simulation is being used to help protect the livestock industries from foreign animal diseases through the development of a national modelling capability to study disease threats. The National Model for Emerging Livestock Disease Threats (NMELDT), developed by Miron et al. in (Miron, et al., 2009), simulates the spread of important livestock diseases, such as foot and mouth disease (FMD), on a national scale by taking into account regional and seasonal factors, different species and production sectors, and marketing systems. The simulation uses a range of inputs such as geography, climatic data, disease life-cycle parameters, and livestock movement data through the National Livestock Identification System (NLIS). The NLIS is Australia's system for identifying and tracking beef and dairy cattle (Australia, 2009). It is a permanent, whole-of-life identification system which aims to ensure that individual animals can be tracked from property of birth to slaughter for bio-security, food safety, product integrity and market access purposes. The NLIS uses individual electronic devices —machine-readable, radio frequency identification devices (RFIDs) — that enable cattle to be permanently identified throughout their lifetime. Each device contains a microchip encoded with a unique number linked to a farm identifier. The animal movement data available in the NLIS database is an important input into the simulation platform.