Agent-based Models

Rule-based, agent-based model

Agent-based models  in which the behaviour of agents is determined by if-then rules are inherently declarative. 

One additional virtue of rule-based, agent-based models is that textual outputs from the models can be linked to the rules that produced specific outputs and these rules can in turn be linked back to documents such as scenario elements and those can be linked back to the evidence that informed them so that there is an "audit trail" or documented "provenance" of the model output going back to any evidence that supports the model-based descriptions of the relevant social processes.

A second virtue is that there is a rule engine that determines when rules can fire and organises the order in which they fire.  As a result, the modeller does not have decide where to insert a new if-then statement into a file of programming code.  This virtue is important since a policy model intended to capture a wide range of circumstances and a perhaps growing number of identified, possible actions by different stakeholders will become increasingly complicated over the course of an OCOPOMO process.

Document Actions

Comments (0)