This post is part of the series Prototyping the Roster for an Understaffed Air Navigation Service Provider: Part1 Part 2 Part 3 Part 4
Quickly recapping the first three parts of our story - we started a new contract with a new-to-launch ANSP, and our mission was to prototype their new ATCO roster. In few regular monthly meetings, we progressed from knowing our customer and its needs, setting up a dedicated SkyRoster Cloud Server, to incorporating all the details of the ANSP and create the virtual image on the SkyRoster platform. As we progressed, we tried to incorporate as much as possible from the customers' feedback at each step, so that we could maximize the progress with each new meeting.
In this part, we are going to talk about the outcome of the hard work done in the first three parts, more exactly about the first two simulations. Then, we will present you the conclusions and what were the next steps to be implemented.
ATCO Rostering Automation, step-by-step
1. Drafting the first two simulations
As stated in the previous article, after discussions with our clients, we decided together to run the first two simulations for the periods with the highest and lowest demand. In this specific case, the traffic peak is during the summer months and the lowest demand is during the wintertime.
From the Sector Opening Time Table provided by our clients, we were able to infer the manpower requirements for the whole year. The SOT provided valuable data regarding how many sectors were opened in the past two years on a daily basis, every hour. So, we were able to identify then the minimum and maximum sectors opened at any time on the day and observe what was the average number of sectors during the whole period. Starting with the nighttime, until early in the morning, there were two sectors opened on average, increasing to five by 8:00 AM until early afternoon, and decreasing gradually to two, reaching a minimum at around 10:00 PM.
Having the minimum, maximum and average number of sectors, it was an easy task to infer the same data for the number of employees, as per each sector opened there was a need for three ATCOs, and two supervisors to manage the whole team. The MPR was then created by observing what was the average number of employees necessary for a shift and considering also the maximum number for the peak period.
The inputs for the first simulation were the following:
nine teams as presented, with 81 employees.
the MANROO pattern: Morning - Afternoon - Night - Rest - Off - Off.
roster timeframe – one month.
As expected for an understaffed ANSP, at the highest demand, the available manpower was insufficient to cover all needed shifts. The best results that we got were still around 540 unallocated shifts for one month in the middle of the highest demand period.
For the winter period, the situation was much better. Using the same inputs and a slightly lower manpower, all shifts were successfully covered by the 81 employees.
2. First conclusions
As the results of the first simulations were ready, we were able to draw conclusions for our clients and prepare recommendations for what was possible to do next. We organized a meeting and presented our analysis and conclusions.
For the summer period, we concluded that the demand can be covered by around seven overtime shifts for each employee during a month. For the winter period, as all shifts were covered, the conclusion was that the scenario could also accommodate training periods in rotation, for every employee.
3. Ways to handle understaffing
There aren't too many options to handle understaffing, especially in a highly regulated industry like Air Traffic Control. The shift schedule for Air Traffic Controllers has to comply with strict rules regarding fatigue management- ICAO Fatigue Management and EASA to name just a few.
In order to comply with these recommendations regarding maximum consecutive shifts and minimum rest time, we were dealing with the unallocated shifts as follows:
in addition to the initially allocated shifts, an Air Traffic Controller can be assigned to work on other shifts during a month, in order to cover the demand.
all these additional shifts are counted as overtime hours.
Morning and Afternoon shifts can be assigned on any of the two Off shifts at the end of the shift cycle (MANROO).
Night shifts can only be assigned on the first Off shift and the second Off shift will be transformed into a Rest shift. By doing this, it was possible to fulfill the overtime requirements and to still comply with the rules and regulations regarding ATCOs’ stress and fatigue risks.
4. Dinner with family
One of the main issues when having to work in shifts is the time spent with the family. A special request coming from our clients was captured after the first two simulations. Being home for dinner is a special moment for everyone, so together with the representatives from the ANSP, we tried to find solutions and create more simulations based on the duration of the Nightshift.
Our clients requested a similar simulation for a Night shift of ten hours, considering that the extra hour could be taken from both the morning and the afternoon shift, so the ATCOs can leave a bit earlier to get home and have dinner with their families.
So, one of the next steps that we needed to do was to reproduce the same scenario for a month and compare the results to find what were the changes in terms of overtime and fatigue risks. The main challenge was to obtain again the minimum possible number of unallocated shifts and to assign the remaining shifts as overtime shifts.
5. If you can't measure it, you can't improve it
After finishing with allocating overtime shifts, our next duty was to measure the number of overtime hours per month/per employee, to observe what is the total number of shifts and hours worked by each ATCO during a month, and to find ways to improve these scores.
Depending on the number of leaves, some of the ATCOs were doing more overtime shifts than the others, but the initial conclusion as can be seen in the Power BI char above was that the maximum number of overtime shifts was six. The representation shows for each ATCO how many overtime shifts were done in a month, and what kind of shifts they had to work in.
In terms of happiness level and fatigue associated risks, we were also able to show for each ATCO the percentage of night shifts done on a monthly basis, considering an average of 4.5 shifts, so the supervisors and managers know which employees were exposed more or less to fatigue associated risks.
In the next part of the story, we will dive into other simulations based on different manpower requirements and rostering periods. Stay tuned for more. Until then, you might consider subscribing to our newsletter, starting your ATCO Rostering Automation journey by signing up for our Free Plan or Booking a Demo with us.
Comments