Policy and Complex Systems
Section 1 , as interaction takes place in the ABM , agents / nodes change their attitudes , features , and position in the network . The fusion we have suggested so far let us investigate what happens to the stability of the network when agents , change , eventually disappear and are replaced by new nodes . 10 For instance , De Caux et al . ( 2014 ) show that for given value of some critical parameter ( in their case , movement ability and age of agents ) the number of separate clusters in the network decreases sharply and generates mega cluster . 11 Such transitions carry important implications for the properties of the networks , such as their resilience to random shocks , which are of crucial importance to policymaker .
As knowledge in this field accumulates , it might be thought that policymaker could fine tune efficiency and stability . In the same sense , but more generally , agent-based network gives insights into the evolution of network statistics over time and on their possible evanescence therefore overcoming some of the problems raised by Edmonds and Chattoe ( 2005 ). Among those , thanks to the transparency of the formation process and of the status of the network nodes , there is the possibility of matching the conventional network measures with a more “ customized ” analysis that can grasp the actual conditions of groups of nodes that are of particular interest ( regions , coalitions , productive sectors ).
Finally , the technique of reverse engineering that we are starting to tackle in the paper is likewise useful in order to diminish the knowledge that policymakers must acquire in order to act .
6 . Concluding Remarks
Complexity economics is currently
facing the challenge of developing theory and tools that can support decision systems in policymaking . Agentbased modeling plays a crucial in completing this task . ABMs can be useful both in deciding ( policymaker level ) and in empowering the capabilities of people in evaluating the effectiveness of policies ( citizen level ). Consequently , the class of ABMs for policymaking needs to be both quite simple in its structure and highly sophisticated in its outcomes . As we have shown , the application of NA to the emergent results can facilitate the achievement of this task by emphasizing the consequences of choices and decisions on the structure of society .
In order to demonstrate the benefits of the matching between ABM and NA we introduce a simple model — recipeWorld — in which networks emerge because of meaningful economic behavior . We then discuss the implications of the joint use of the two techniques at length , focusing on the role of dynamic network models in policymaking and by introducing a research challenge that we are undertaking . Since it is easier to have network data ( i . e ., social network data ) than detailed behavioral individual information , we can try to understand the relationship between the dynamic changes of the networks emerging from ABMs and the behavior of the agents . As we understand these relationships , we can apply them to actual networks , trying to understand the content of the behavioral black boxes of realworld agents .
References
Aumann R . J . & Mayerson , R . B . ( 1988 ). Endogenous formation of links between players and coalitions : An application of the “ Shapley Value ”. In the Shapley Value , A . Roth Ed ., Cambridge University Press , Cambridge .
89