Enhancing ABM into an Inevitable Tool for Policy Analysis
tion of this mathematical method is that it does not facilitate communication with domain experts because mathematical equations are not understandable to everyone . Also , NEM does not help to gain insights into individual behavior and decision making , as it describes the system from a macro perspective . In terms of decision support , NEM can be used to categorize and distinguish policies through mathematical equations . However , since certainty is one of the assumptions in this approach , the answer to what-if scenarios using this approach are not always reliable . This also holds for tracking behaviors because not only the system is analyzed at macrolevel , individuals are considered fully rational with complete information . These assumptions may also not justify answers for certain behaviors and reactions when monitoring the implemented policies .
Unlike NEM , traditional game theory takes an individual-based approach , which provides a means of specifying the population in the problem definition . However , TGT is about outcomes rather than individual behavior . Therefore , it is not the most suitable tool for tracking individual behaviors and reactions for selecting a policy or monitoring an implemented one . While TGT can be used to define evaluation measures and identify extreme values and worst case scenarios through computing equilibria , as a standalone tool , it has other limitations . For example , it does not provide a test bed for participatory decision making unless it is used with serious gaming ( SG ). TGT does not support the identification of policy alternatives either : there is no way of identifying the attributes for policy alternatives and it is infeasible to link policies to evaluation measures .
System dynamics ( SD ) is a computer simulation approach that makes use of differential equations . Therefore , all the benefits of mathematical descriptions , such as formulating policies and their attributes and defining evaluation measures , are facilitated with SD . Owing to the availability of tools , displaying and presenting various policies is practical . In addition , SD like other simulation approaches enables tracking of system behavior . However , the major limitation of SD is that the system is not viewed as a collection of individuals . Therefore , it is infeasible to gain insights into populations and the decision-making behavior of individuals and thus not possible to track behaviors and reactions at individual level toward a policy . Nonetheless , the general processes and outcomes are traceable . Furthermore , instead of identifying boundaries and resources for a policy problem , to make a system dynamics simulation , these aspects need to be defined beforehand .
Serious gaming ( SG ) is one of the most useful tools to define a policy problem . It also facilitates the definition of evaluation measures and identification of their link with the problem definition . When the players become involved in a serious game , they are able to place themselves in the situation to find out how policies would affect them . This facilitates the identification of the association between policy alternatives and evaluation measures . The limitations of serious gaming ( SG ) are all related to the limited number of players in a game which is normally much smaller than the number of agents in the system in which the policy would actually be implemented . This limitation makes it difficult to :
– rely on the results for what if scenarios , – test extreme values and worst case scenarios , and – track reactions toward policies before and after a policy implementation .
69