An agent-based framework for modelling social activities and travel Nicole Ronald, Theo Arentze, and Harry Timmermans

Overview What are we modelling and why? How agent-based modelling is useful for systems with social and spatial components Designing a model of social activities and travel behaviour Discussion and conclusions

What are we modelling and why?

How agent-based modelling is useful for systems with social and spatial components

Designing a model of social activities and travel behaviour

Discussion and conclusions

What are we modelling?

Social/spatial

Framework Selection Influence Activity generation who with, how often, where

Selection

Influence

Activity generation

who with, how often, where

Agent-based modelling A computational method that enables a researcher to create, analyse, and experiment with models composed of agents that interact within an environment (Gilbert, 2008) Complex, nonlinear, discontinuous, or discrete interactions Heterogeneous population Agents move around the environment Complex behaviours required, including learning and adaptation

A computational method that enables a researcher to create, analyse, and experiment with models composed of agents that interact within an environment (Gilbert, 2008)

Complex, nonlinear, discontinuous, or discrete interactions

Heterogeneous population

Agents move around the environment

Complex behaviours required, including learning and adaptation

Model design Maintain relationships Share experiences Share information Learn about environment Shall we try the new café on the river? Sounds good. How about Thursday? No, Friday is better for me. That’s good for me too. 7pm? Perfect -- see you then! host respondent

Services determining who to un/form a connection with keeping a schedule determining whether to initiate a discussion about an activity or send an invite evaluating invite/activity proposals undertaking activities evaluating activities, in terms of satisfaction with location, people and time updating own properties sharing information about the environment with others participating in a club, i.e., deciding to join/leave

determining who to un/form a connection with

keeping a schedule

determining whether to initiate a discussion about an activity or send an invite

evaluating invite/activity proposals

undertaking activities

evaluating activities, in terms of satisfaction with location, people and time

updating own properties

sharing information about the environment with others

participating in a club, i.e., deciding to join/leave

Interaction

Discussion Caters for all components of our framework (selection, influence, activity generation) Novel use of interaction/negotiation theory combines multi-issue negotiation and scheduling combines behavioural choice and interaction theory Validation methodology Non-dynamic environment, dynamic relationships

Caters for all components of our framework (selection, influence, activity generation)

Novel use of interaction/negotiation theory

combines multi-issue negotiation and scheduling

combines behavioural choice and interaction theory

Validation methodology

Non-dynamic environment, dynamic relationships

Conclusions ABM is a suitable method for modelling social activities Novel use of interaction theory for negotiating about activities Interaction prototype in experimental phase, complete model in implementation

ABM is a suitable method for modelling social activities

Novel use of interaction theory for negotiating about activities

Interaction prototype in experimental phase, complete model in implementation

Questions? Nicole Ronald [email_address]

Nicole Ronald

[email_address]