Case studies
Simulation: 120 ATCOs, 37 overtime shifts
What changes when a 96-controller unit adds 24 staff? This April 2022 simulation used a 120-person workforce to examine coverage and overtime against the specified demand.
The result is a simulation with increased staffing, not an observed saving attributable to software alone. Staff were represented by anonymised identifiers in the published analysis.
Scenario setup
| Input | Value |
|---|---|
| Strategic roster | 1–30 April 2022 |
| Tactical period examined | 11–17 April 2022 |
| Workforce | 120 employees in 12 teams of ten |
| Pattern | Morning–Afternoon–Night–Rest–Off–Off (MANROO) |
| Morning duty | 07:00–14:00 |
| Afternoon duty | 14:00–21:00 |
| Night duty | 21:00–07:00 |
Annual leave and other duties were distributed for the simulation using annual averages. The monthly roster contained 488 days of leave and training: 326 annual-leave days and 162 English-training days.
The staffing requirements were 15 ATC positions for each morning and afternoon shift, 12 at night, and two supervisor positions on each shift.
Tactical allocation assumptions
The model included executive (EXE), planner (PLN), supervisor (SUP) and air flow manager (AFM) positions.
EXE and PLN assignments allowed a maximum of two different positions, one hour each, followed by at least an hour’s break. The model allowed SUP position work for 100% of morning and afternoon duties and 50% of night duties. AFM coverage ran from 07:00 to 18:00.
These percentages describe the project’s configuration, not universal ATC break requirements. Strategic and tactical rosters were generated using constraint optimisation, with outputs analysed in Power BI.
Overtime result
The model covered the specified staffing requirements with 37 overtime shifts: 19 morning and 18 afternoon shifts. No night overtime was required in this scenario.
The 37 overtime duties were assigned to 37 employees, one each. Team H had no overtime assignments. Regular morning and afternoon duties were approximately balanced; differences between teams reflected the modelled availability and leave.
Reported time on position
| Period | Reported time on position |
|---|---|
| Morning EXE/PLN | 57.14% or 71.42% |
| Morning SUP and AFM | 100% |
| Afternoon EXE/PLN | 42.85%–71.42% |
| Afternoon SUP | 100% |
| Afternoon AFM | 57.14% |
| Night assignments | Approximately 46%–54%, around the 50% target |
The source’s monthly hours summary ranged from 14 to 127 hours for the identified employees. Four employees at the lower end each had 25 consecutive annual-leave days. Its team position-time summary reported minimum, average and maximum values of 58, 74 and 98 hours respectively.
The reported team shares of time on position ranged from 55.85% to 61.10% for EXE/PLN, and from 57% to 64% when SUP and AFM were included. Staff holding multiple qualifications could appear across more than one position category.
Interpretation
Adding 24 staff changed the capacity available to cover demand. The model also included substantial leave and training, so those assumptions must accompany the overtime result.
The related 96-controller simulation used May rather than April and different leave assumptions. The two scenarios are informative but are not a controlled comparison of headcount alone.
The original article used rounded summary percentages and an ambiguous “efficiency factor”. The position-time ranges above retain the reported measures without treating that label as a separate performance result.
See case studies and methodology for further context.