PBS Core Algorithms
To achieve the highest overall satisfaction, AOS implements the Global Optimization to our PBS core.
Before awarding, the AOS PBS core generates a huge number of preferred lines for all crewmembers, and selects a solution by global optimization which maximizes the overall crew satisfaction and company goals. Then, the core generates multiple best legal lines for each crewmember, starting from the most senior, and uses this global solution as a guideline to select and award the best fit among all best legal lines. The awarded pairings are then removed from the available pairings.
The Global Optimization runs several times at different seniority levels to guarantee it always has the best guideline for each crewmember when awarding.
Unlike the greedy algorithm, which awards an arbitrary best line, global optimization allows the core to evaluate all best lines and awards the one with the least impact to junior crewmembers and company goals.
Because of the use of global optimization, it constantly monitors and predicts when is the best time to begin to generate the best lines that contain the pairings needed to be covered at any given date; therefore, the un-stacking approach is not necessary at all.
The real spirit of using global optimization is to find the best fit among the best lines. It is AOS' true strength, while other PBS vendors use greedy and swapping based algorithms. No other traditional approach comes even close to the high overall satisfaction we can achieve with AOS' PBS.