|
?
If an optimizer isn't providing the solution instantly, there is definitely
room for improvement. Solving a problem within minutes instead of hours
will have a direct impact on the planning process, along with the ability
to react to changes. Of course, increased performance is also needed to
be able to solve more comprehensive planning problems.
There is always great potential from increasing the scope. Let’s
look at crew planning as an example. Deciding which crew member shall
operate which flights is today done by first creating anonymous crew blocks,
also called pairings. After this the pairings are combined into individual
crew rosters. The pairing step exists because historically the problem
was too big to solve in one step. However, this is not the case anymore.
For the integrated crew planning problem the potential for improvements
in cost savings and in reduced process time is much larger compared to
the sequential approach. Within a few years from now I believe that separate
planning of pairings and rosters will be considered old-fashioned.
Yes, and there are also other potential areas, one example is planning
crew and vehicles together. In practice the crew members and the vehicles
must be at least synchronized for critical connections, but the planning
of them is usually separated. Our experience both from real operation
and from experiments show that synchronizing the planning of aircraft
and crew makes it possible to reduce the crew cost. More importantly,
it also makes it possible to increase the operational stability by eliminating
the majority of the critical aircraft changes.
Flexibility! All our products support our legality language Carmen Rave.
This provides our clients the necessary flexibility to cope with long-term
process changes as well as short-term changes in planning practices. Today
our clients actively use this tool to calibrate both the daily production
and also to perform simulations. The feedback from our clients is that
they achieve an additional average saving of around 2% as a result from
the ability to constantly adjust properties like legality, quality and
detailed costs.
One particular area that I think will be of great importance is the introduction
of roster analysis and roster repair functionality. This will now be available
in a number of additional functionality packages which allow rosters to
be repaired, compared and re-optimized. The ability to analyze the difference
between the original plans and the final outcome collectively with the
tools for re-optimizing already published rosters will make it possible
to dramatically reduce the number of problems that reach the day of operation. |