2007 Symposium
Computational Science and Public Policy
May 24, 2007
Computational Science and Public Policy
May 24, 2007
"Computational Methods as an Aid for Public Policy Formulation"
Robust public policy must account for the behavior of diverse actors. Most policy models draw on static, highly simplified representations of behavior. Yet unanticipated and un-accounted for behavior often determines the success or failure of implemented policies and programs.
Computational methods used in scientific analysis of complex adaptive systems were the focus of this symposium. These “agent-based” simulations, constructed of many autonomous interacting software objects (“agents”) representing the different actors, are used to generate ensembles of potential evolutionary scenarios. Policy analysts can adjust constraints, incentive structures, and rules in these simulations and observe the dynamics of adaptive behavior across the actors. These simulations allow for full transparency into the roles, rules, goals, and relationships between the modeled actors – allowing for a rigorous, repeatable, and falsifiable framework for policy analysis.
Some specific areas where these methods are being applied were presented along with discussion of the challenges and disadvantages of using agent-based simulation for policy analysis.
SPEAKERS:
- Rob Axtell, Professor, George Mason University
- Darby Jack, Sr. Fellow, Columbia University
- Edward MacKerrow, Los Alamos Labs
- Jonathan Schneider, Atalaya Institute
Robust public policy must account for the behavior of diverse actors. Most policy models draw on static, highly simplified representations of behavior. Yet unanticipated and un-accounted for behavior often determines the success or failure of implemented policies and programs.
Computational methods used in scientific analysis of complex adaptive systems were the focus of this symposium. These “agent-based” simulations, constructed of many autonomous interacting software objects (“agents”) representing the different actors, are used to generate ensembles of potential evolutionary scenarios. Policy analysts can adjust constraints, incentive structures, and rules in these simulations and observe the dynamics of adaptive behavior across the actors. These simulations allow for full transparency into the roles, rules, goals, and relationships between the modeled actors – allowing for a rigorous, repeatable, and falsifiable framework for policy analysis.
Some specific areas where these methods are being applied were presented along with discussion of the challenges and disadvantages of using agent-based simulation for policy analysis.
SPEAKERS:
- Rob Axtell, Professor, George Mason University
- Darby Jack, Sr. Fellow, Columbia University
- Edward MacKerrow, Los Alamos Labs
- Jonathan Schneider, Atalaya Institute