Magnus Ek is a relatively recent addition to the Jeppesen Scheduling Safety team, but he brings a wealth of experience, boasting almost 25 years of expertise in crew planning. With a particular focus on pairing and rostering optimization, Magnus below sheds some light on the concept of an 'objective function.'
-Crew planning optimizers aim to generate 'solutions' that are both a) legal and feasible, and b) as 'good as possible'. The former involves adhering to regulations, labor agreements, and various internal scheduling rules essential for a practical plan. Factors like the number of captains at each base and their availability over the planning period are also crucial considerations.
-The latter, achieving the best possible outcome, is where the objective function plays a role. 'As good as possible' is quantified by assigning a value to everything crucial for the operator. This could be a 'penalty' for not granting a bid from Captain Svensson to have the 14th off or a penalty for having a small margin to a maximum Flight Duty Period (FDP) rule, which could lead to operational instability if flights are delayed. Quality aspects, such as evenly distributing standby duties among the crew, also contribute to cost components in the objective function. Additionally, when connected to a BMM like BAM, there's even a cost associated with planning flights where the crew is predicted to be low in alertness.
-During the optimization process, which may span many hours for a large operator, the optimizer alternately generates legal pairing and roster 'candidates' and evaluates the full solution using the objective function. By alternating between generation and evaluation, the optimization process cleverly navigates toward progressively more attractive planning solutions—those with the lowest 'cost' in the objective function. If something appears undesirable in a solution, it's simply a matter of formulating another penalty and refining the plan. It's a fantastic platform for continuous improvement and running what-if scenarios to quantify the impact of potential changes, ultimately steering clear of bad decisions and pursuing most favorable outcomes.
Well explained, thank you Magnus! Read more about Jeppesen Crew Pairing and Rostering here.