Agency: A Framework for Modeling Decisions

An initial model for simulating decisions, learning, and consequences.

Version 1

This conundrum tackles a foundational problem in systems thinking: how do you model a world full of decision-makers who see things differently, want different outcomes, and act under uncertainty?

Classic system dynamics is powerful for understanding flows, feedback, and structure—but it struggles with agency. Real institutions, governments, and individuals don’t behave like mechanical variables; they choose, interpret, react, and learn. Without a way to represent those internal processes, models of complex systems miss the very forces that drive them.

In this project, I develop a framework that brings agency into system dynamics by breaking decisions into four parts: Objectives, Resources, Perceptions, and Actions. The result is a universal architecture for simulating how agents think, what they value, what they control, and how they respond to a changing world.