A School of Emergent Priorities research poster builds and deploys a round-robin scheduler for a farmers market's single shared card reader, and finds equal device-time has no bearing on customer queue length.
Fourteen Saturdays of hand-off logs hold every stallholder's turn equal, and the customer queue drifts regardless
A market’s single shared card reader has never needed a schedule; stallholders simply asked. A research poster published today by the School of Emergent Priorities replaces the asking with an algorithm, and reports what the algorithm chooses to be fair about.
Dr Runa Adegoke, Senior Lecturer and Convenor of the Strategic Drift Survey, led the build with Dr Osei Vandermeer, Research Fellow and Convenor of the Review Cadence Observatory, School of Continuous Improvement. The pair instrumented a fortnightly market’s handheld reader with hand-off logging across a full trading season, then set a weighted round-robin controller to hold every stallholder’s share of the device to an equal split, irrespective of how many customers were actually waiting at each stall.
The reader was always going to be fair to somebody. We hadn’t decided yet whether that was the stallholder or the queue.
— Dr Runa Adegoke, Senior Lecturer and Convenor, Strategic Drift Survey
“Every scheduler we build ends up teaching us which measurement we left out,” said Dr Vandermeer, who plans to fold footfall counts into the controller’s next revision. The University counts the deployment among its clearer examples yet of a governance instrument built and audited in the same season.
The full poster is available from the University’s research repository under an open licence, doi:10.5555/slop.cmzmgb.