Exploring Wealth Distribution Through Sugarscape

Jordan Albright

Purpose

• Analyze the wealth distribution of agents in Sugarscape over time

• Compare calculations with other studies

• Assess validity of project in scope of social science

Similar Studies

• An Agent-Based Model of Wealth Distribution

• Impullitti and Rebmann, 2002• Studied sugarscape wealth distribution in

context of classical and neo-classical economics

• Found it difficult to bridge model and reality in agent based modeling

Similar Studies (cont.)

• Aligning Simulation Models: A Case Study and Results

• Axtell, Axelrod, Epstein, Cohen, 1996

• Attempted to synthesize results from Sugarscape and Axelrod model

• Difficult to streamline results of models when given same initial conditions

Sugarscape

• Movement– Vision

• Reproduction• Death• Sugar regrowth

Wealth Distribution Analysis

• Lorenz Curve– Proportion of

distribution

– Line of perfect equality

– Basis for many other functions

Wealth Distribution Analysis

• Gini coefficient– Measure of inequality

– Ratio ranging from 0 to 1

– Typical range of developed countries: .3 to .4

User Controls

• User can specify:– Initial population– Birth energy– Max. metabolism and vision– Inheritance on or off– Display vision of agents

Results

• Comparison of results with Impullitti:– Gini coefficient in project

• Without inheritance: .35 to .4• With inheritance: .4 to .5

– Gini coefficient in Impullitti: .3 to .33• Different rules for agents• Different initial conditions

Results (cont.)

• Non-inherited

• Small fluctuations in GC, vision, metabolism

Results (cont.)

• Inherited

• Metabolism approaches lower limit

• Vision increases but does not reach maximum

Results (cont.)

• Based on visual representation, bottlenecking

• Regions of unequal wealth distribution