The Loo Loop Ethical AI Agent: A Queuing- Theoretical Model of Mixed-Gender Toilet Allocation Under Equity Constraints

Pascal Stiefenhofer

We develop a queuing-theoretical model for mixed-gender public toilet systems, incorporating contemporary Equity, Diversity, and Inclusion (EDI) principles. The model uses a multi-server Markovian framework with gender-mixed arrival streams, incorporating user discomfort and strategic behavior such as balking and reneging. We introduce the Flush Equity Index (FEI), a novel fairness metric designed to evaluate equity in access across gender and identity groups. Our analysis reveals paradoxes inherent in socially progressive restroom designs: while increased equality of access aims to reduce inequities, it may unintentionally decrease system efficiency and user satisfaction. To address this, we propose several policy interventions, including the Privacy-Aware Stall Assignment Protocol (PASAP), behavioral nudges, and a blockchain-based ToiletCoin system. The ToiletCoin system optimizes queue allocation by rewarding users for cooperative behavior, such as nudging and prioritizing high-urgency users, while reducing discomfort. Finally, we recommend using the FEI in restroom equity audits to guide design and policy decisions.
PDF