Presenter/Author Information

Claudia Pahl-Wostl
Eva Ebenhöh

Keywords

heuristics, modelling human behaviour, experimental economics, cooperation, fairness

Start Date

1-7-2004 12:00 AM

Abstract

Human behaviour is one of the key factors to understand the causes for common pool resourceproblems and to develop policies to promote more sustainable resource management regimes. Agent basedmodels can help to investigate the role of important processes in this respect such as factors determining thedegree of trust and cooperation in a group. We have chosen a pragmatic approach to represent humanbehaviour by assuming that agents can be characterised by a set of attributes and their behaviour can bedescribed by a set of simple decision heuristics. Individual agents differ in their importance of attributes (e.g.fairness, cooperativeness, trust), in their rules how to choose a heuristic, and in their responses to socialinteractions. The assumptions are tested by using data from experimental economics describing the behaviourof players in simple games dealing with resource allocation. A set of specific attributes and heuristics wasderived by analysing data from different games. The plausibility and generality of the behavioural model istested by applying it to different data sets from different games. We expect from these simulations insightsinto behavioural patterns that determine processes of social learning and negotiation. The modelling approachwill be applied and tested with data from case studies where actors make decisions in a real world context ofdealing with a resource management problem.

Share

COinS
 
Jul 1st, 12:00 AM

Heuristics to characterise human behaviour in agent based models

Human behaviour is one of the key factors to understand the causes for common pool resourceproblems and to develop policies to promote more sustainable resource management regimes. Agent basedmodels can help to investigate the role of important processes in this respect such as factors determining thedegree of trust and cooperation in a group. We have chosen a pragmatic approach to represent humanbehaviour by assuming that agents can be characterised by a set of attributes and their behaviour can bedescribed by a set of simple decision heuristics. Individual agents differ in their importance of attributes (e.g.fairness, cooperativeness, trust), in their rules how to choose a heuristic, and in their responses to socialinteractions. The assumptions are tested by using data from experimental economics describing the behaviourof players in simple games dealing with resource allocation. A set of specific attributes and heuristics wasderived by analysing data from different games. The plausibility and generality of the behavioural model istested by applying it to different data sets from different games. We expect from these simulations insightsinto behavioural patterns that determine processes of social learning and negotiation. The modelling approachwill be applied and tested with data from case studies where actors make decisions in a real world context ofdealing with a resource management problem.