 |
FriarTuck provided a workforce scheduling system (an adapted
version of FTRosterer) to schedule scientists and engineers for the Mars
Exploration Rover mission. In Science team alone, there were about 285
scientists working on the Mars Exploration Rovers mission, filling 22
different operational roles. Any one scientist is qualified to fill one
to five different roles based on their expertise and experience.
Creating and maintaining the staffing schedule for the mission requires
that the duty planner staff the 22 roles for both rovers while adhering
to rules about rest time and work weeks that are designed to minimize
personnel fatigue. |
On top of this, the whole mission is run on Mars time. Since a Mars day is about 40 minutes
longer than an Earth day, this means that staff show up for their shifts 40 minutes later
each day. Furthermore, there were another 4 teams in addition to the Science Team. It's
an extremely challenging problem to find an optimal schedule that works within all these
constraints. To estimate the complexity, the problem size comprised 285 staff x 150 Martian
sols x 38 Positions x 2 Rovers leading to over 3 million binary decision variables.
The MER optimization criteria outlined:
- Minimize fatigue through patterns: 4–2/4–3/4–4
- Maximize the fulfillment of preferences
- Maximize the appropriate assignment of scientists to positions
- Balance assignments of scientist for fair schedule
- Maximize the rotation of staff across different positions
Impact on the Mission: Such optimized staff scheduling for the mission was considered
crucial because- Shift cycle length and patterns affects staff fatigue/performance;
Frequency of staff in primary skill position affects service quality; Fairness in scheduling competing
scientists affects motivation; Frequency of shift rotation affects learning
experience; Staff preferences affect motivation and performance and finally the “Schedule” affects
the ability to flexibly re-allocate staff in response to malfunctioning or
damaged equipment, unforeseen discoveries etc.
“FTRosterer has enabled us to check existing schedules for violations and
ensure that no staff member is double-booked or overworked. It has provided a means
of easily managing the schedule through convenient data entry screens and options
for export of the roster to Excel and CIP formats. Without FTRosterer, the duty planners
would have to manually integrate staffing plans from ad hoc sources such as spreadsheets
-- a very labor intensive and time-consuming process. Checking for constraint violations
and double-booking of personnel would be extremely difficult. With FTRosterer, the
integration and constraint-checking tasks are automated and efficient”
Joan D. Walton, NASA Ames Research Center. |