Under the Rostering Automation Prototyping Services contracts, we had the chance to model exciting and varied rostering simulations that could provide valuable learnings for other ANSPs. We will present some of these learnings in a series of Case Studies based on the actual simulations that we have performed for our customers, but using 100% anonymised data.
1. Problem description
An ATC Unit with 96 ATCOs rostered in teams on a fixed shift pattern sees significant variations in terms of overtime distribution per individual staff, and increased staff leakage during the night shifts.
How can the roster balance overtime shifts between available employees for a 1 month period (April 2022) and minimise leakage?
2. Initial setup
The data used for this modelling was uploaded to the Customer's server on the SkyRoster Cloud.
All annual leave days and duty days (office work, training, English courses, etc.) were considered by inputting them into the system for the simulated period (distributed randomly per employee, averaged out per year).
As a preliminary step, the roster for April 2022 was generated using the entire available staff in the ATC Unit, using a Team-based roster with a Fixed Shift Rotation Strategy.
Type of roster: Team-based.
Roster period: 1st - 30th of April 2022.
Shift Allocation: Fixed Shift Pattern.
Pattern: MANROO (Morning-Afternoon-Nigth-Rest-Off-Off).
Shift configuration:
Morning shift: 07:00 – 14:00.
Afternoon shift: 14:00 – 21:00.
Night shift: 21:00 – 07:00.
Staff: 120 employees grouped equally in 12 teams, 10 ATCOs per team (24 additional ATCOs).
Manpower Requirements (MPR):
The MPR was updated to accommodate all the necessary positions according to the Sector Opening Timetable and was defined using the Air Traffic Control Officer (ATC) and Supervisor (SUP) qualifications as follows:
For ATC:
Morning: 15 positions.
Afternoon: 15 positions.
Night: 12 positions.
For SUP
Morning: 2 positions.
Afternoon: 2 positions.
Night: 2 positions.
The next step was to generate the tactical roster and solve it for a week. The week chosen for this analysis is 11 – 17 of April 2022 and the Sector Opening Timetable was the following:
The necessary positions to fill were Executive (EXE), Planner (PLN), Superior (SUP) and Air Flow Manager (AFM). The procedure by which the staff was assigned to the Tactical Roster was the following:
Maximum of two different worked positions, one hour each, followed by a minimum of an hour break in case of EXE and PLN.
The employees in the SUP position can perform 100% during the day shifts (M and A) and 50% during the night (N).
The AFM position is open between 07:00 and 18:00.
Both strategic and tactical rosters were generated using the SkyRoster AI-powered constraint-based scheduling engine. Afterwards, the data was imported into Power BI to observe the shifts distribution per employee and team, the overtime distribution, the time spent on position and the staff leakage.
3. Statistical results
April 2022 Strategic Roster
The SkyRoster application has accommodated the entire MPR using 37 overtime shifts: 19 Morning Shifts and 18 Afternoon Shifts.
The strategic roster includes 488 leave days: 326 Annual Leaves and 162 English Training days.
April 2022 Tactical Roster
The SkyRoster Engine assigned all positions successfully, respecting the resting time specific to each position.
4. Analysis
April 2022 Strategic Roster
It can be observed in Table 1 that the number of Morning and Afternoon shifts are almost equally distributed. The lower result for the Night shifts is due to the lower number of needed shifts in MPR for the night time. The same result can be observed for overtime shifts which are almost equally distributed between Morning and Afternoon shifts. A major improvement over the Excel roster: there is no overtime needed for the Night shift!
From Figure 4-1, it can be seen that shifts are almost evenly distributed between the teams, with only a small variation due to leaves affecting employees in certain teams.
The Overtime Distribution figure shows that the overtime shifts in April are evenly assigned to all employees, with only one overtime shift per month. Therefore, the simulation displays 37 employees with overtime shifts. It can be observed that the employees in team H are not performing any overtime shifts.
April 2022 Tactical Roster
Figure 4‑4 shows the employees' schedules for the SUP and the AFM positions. It can be observed that the employees who perform these two shifts work 100% of the time at work, and in Figure 4‑5 that the rest of the employees who work on the EXE and PLN positions spend 71.42% or 57.14% of the time in position; this percentage varies depending on how the system distributes the necessary positions.
Figure 4‑6 below shows the employees' schedules for the SUP and the AFM positions. Same as in the case of Morning Shifts, the employees who work as SUP work 100% of the time. Meanwhile, the employees who work in the AFM work 57.14% of the time. The employees who work in the EXE and PLN positions spend between 42.85% and 71.42% of their time in positions (see Figure 4‑7).
The percentage of Night shifts is the closest to the standard of 50%, varying from 46% to 54%. Examples of night shift positions can be observed in Figures 4‑8 and 4‑9 below.
Figure 4-10 shows the distribution of time at work (in hours) in the entire month. It can be observed that the maximum is 127 hours for the employee TBE1 Team32, and the smallest value is 14 hours for the employees C02, C04, C05 and C23. This can be explained by observing that these employees have 25 days of consecutive annual leave each.
Analysing Figure 4‑11, we can observe the teams' minimum, maximum and average time on position (in hours). The minimum time on position for a team is 58, the average is 74, and the maximum is 98.
Figure 4‑12 shows the percentages of time on position versus breaks for EXE and PLN positions.
Figure 4‑13 is similar to Figure 4‑12, only that it includes the SUP and AFM positions.
In both cases, we can observe that there is no team to work more than 61,10% of the time and less than 55,85% during the entire period.
For the overall statistics (i.e. including SUP and AFM), the teams are not working more than 64% and less than 57%.
A breakdown of each type of working position and breaks is displayed in Figure 4‑14 and Figure 4‑15.
It can be observed that there are employees who work in multiple positions, either SUP or AFM, either mixed (SUP&AFM and EXE & PLN), depending on the system assignation and qualifications.
5. Conclusions
The optimised roster for April 2022 minimises the variation of overtime between teams and individual employees by using only 37 overtime shifts for the entire month.
All leave days are guaranteed for the entire year, with a proportion of them being assigned in April 2022.
In the Tactical Roster, staff work in the Morning and Afternoon shifts at the planned percentage of 66%. The SUP makes the exception, who work 100% of the time in both types of shifts and AFM, who work 100% of the time in the morning and 57% in the afternoon.
In the afternoon, there is a transition from 5 open sectors to 4 open sectors, and the AFM position is closed.
No employee or team works more than 71.43% of the time, except for the SUPs in the Morning and Afternoon Shifts.
The time spent on position during the Night shift is 50% for all three positions: SUP, EXE and PLN.
On average, the efficiency factor is 50% during the night shifts and 33% during the Morning and Afternoon shifts.
6. Full Power BI Dashboard
The fully-interactive Power BI Dashboard with anonymised data for the entire simulation can be found below.
You can download more case studies and have the opportunity to learn from the experiences of other ANSPs who have successfully optimised their rosters using SkyRoster. With this information at your disposal, you can better understand the benefits of implementing SkyRoster in your organization and achieve greater efficiency in your workforce scheduling.
Comments