John Bruntse Larsen: The purpose of the project is to investigate the use of organizational reasoning in multi-agent based simulation.
Examples of organizational reasoning in this context is to be aware of the norms and roles in an organization and to be able to distinguish between personal goals and goals of the organization. AORTA abbreviates "Adding Organizational Reasoning To Agents" and is a multi-agent reasoning framework developed at DTU and TU Delft, The Netherlands. AORTA enriches agents with organizational reasoning, and this project investigates how AORTA can be applied for simulation purposes. The project also attempts to apply such simulation at a real hospital to solve the problem of predicting bottlenecks that can occur due to changes in the schedule.
Simulation in general is advantageous since it can be used to predict what can happen in the real world, and multi-agent based simulation in particular has been shown to work well in environments with complex human interaction. Organizational reasoning is also especially useful in environments with humans as it can model how organizational norms affect the agents and what it means to the system as a whole. Such a normative multi-agent simulation with organizational reasoning has potential to yield simulation tools in which the norms of the real world are introduced into the simulation automatically from real world data with process mining. This would lower the work needed to fit the simulation to a given environment and yield more detailed simulation models automatically over time.
At Danish hospitals, simulation has been used to estimate how big an emergency department needed to be when considering the actual work involved in treating patients and how the people move around the hospital.
PhD project: 2016 -
Supervisor: Jørgen Villadsen, Section for Algorithms, Logic and Graphs