Sugarscape

Sugarscape

Sugarscape refers to a computer simulation model that mimics the actions and interactions of artificial agents situated on a landscape where the primary resource is sugar. It was originally developed by Joshua M. Epstein and Robert Axtell in 1996 and has since become a popular topic of study in the field of agent-based modeling and simulation.From: Multi-Agent Systems [2019], Handbook of Bioinspired Algorithms and Applications [2019], Sensitivity analysis of an agent-based simulation model using reconstructability analysis [2021]

Agents in Social Science

View Chapter

Purchase Book

Published in Mariam Kiran, X-Machines for Agent-Based Modeling, 2017

Mariam Kiran

Sugarscape is also useful to see how societies develop and evolve. If agents were equipped with genetic material, this could be inherited by newborn child agents, allowing best genes to be carried onto new generations. Running such an experiment, “over a period of time, the characteristics of the population of agents converge towards certain traits, namely good vision and a low metabolism” [118].

Sensitivity analysis of an agent-based simulation model using reconstructability analysis

View Article

Journal Information

Published in International Journal of General Systems, 2021

Andey M. Nunes, Martin Zwick, Wayne Wakeland

In agent-based simulation (ABS), agents interact with each other in a dynamic environment. By following simple rules, these interactions result in emergent behavior patterns. SugarScape is a widely studied ABS model developed by Joshua M. Epstein and Robert Axtell (Epstein and Axtell 1996). The NetLogo Wealth Distribution model, developed by Uri Wilensky, is based on the SugarScape model and includes output variables for the Gini coefficient, a measure of wealth inequality, and for the class distribution in the simulation population (Wilensky 1999). This project applied a machine learning methodology to the outputs generated by Wilensky’s Wealth Distribution model to answer the following questions, Can a machine learning algorithm detect relations between model parameters and model output that augment our understanding of the model? Specifically, can such an algorithm reveal the degree to which the model parameters and their interactions predict the model output?

Explore chapters and articles related to this topic

Agents in Social Science

Sensitivity analysis of an agent-based simulation model using reconstructability analysis