Fair-MPC: A Framework for Just Decision-Making

Abstract

While control theory can be pivotal in addressing societal challenges and helping policymakers in shaping our future, control schemes must incorporate elements of social justice to be both up to the task and fair. In turn, this requires the formulation of new constraints and control objectives and their integration into existing or new control design strategies. Devising a formally sound framework to ensure social fairness in control can enable a leap in understanding the implications of justice when closing the loop. In this challenging context, we propose and analyze the properties of Fair model predictive control, an economic model predictive control scheme promoting fairness by integrating into the design strategy social fairness metrics, articulated in equality and equity. This allows optimizing input sequences that balance fairness and efficiency, which are often in conflict.