What agent-based modeling system has the features we would like?

[djinnome]

• How can the syntax and semantics of an ABM model be serialized in a declarative format (XML, JSON, etc)
• Is there a GUI for specifying constraints, and objectives, learning from data and visualizing outputs?
• Does it have a way to calibrate parameters?
• Does it have a way to solve for behaviors that result in optimal outcomes?
• Is there a GUI for eliciting domain knowledge from experts through a process diagram?

[djinnome]

@akhavan12 said in response to question 1:

It is always possible to generate xml or json files from the steps and parameters, here is an example from Mesa Python
library.

   def export_to_xml(self, filename):
        root = ET.Element("Model")
        root.set("current_step", str(self.current_step))
        agents = ET.SubElement(root, "Agents")

        for agent in self.schedule.agents:
            agent_elem = ET.SubElement(agents, "Agent")
            agent_elem.set("unique_id", str(agent.unique_id))
            agent_elem.set("infected", str(agent.infected))

        tree = ET.ElementTree(root)
        tree.write(filename)

[djinnome]

this is useful for tracking simulation outputs, but I was looking for a way to represent the model before we run the simulation