
Workforce scheduling
October-November 2021
About the company
Charter Aviation Services (CAS) is an aviation school that trains commercial pilots en Colombia. They deliver theoretical and practical courses in all aviation related areas. The have their own fleet of training planes and are based in Barrancabermeja airport where they attend students from all around Colombia.

The Challenge

The goal is to help CAS optimize their class schedule. To program a class, three things are required: a student, an available instructor and an available plane. The objective is to fit the maximum number of classes in a limited period of time. So far CAS used to realize the schedules manually. Which is not only suboptimal but also very tedious because of the huge amount of constraints that have to be taken into account. For example, the maximum and minimum daily hours that instructors and students can work, planning the schedules in compliance with the regulations that requires specific maintenance programs for the planes, taking into account setup times when a plane or an instructor is changed between two classes, etc.

11%
19%
3
increase in
programed hours
in schedule variability
new indicators
Winners of the Challenge

Arturo Gonzalez Lian
Arturo's software is based on a set of algorithms that take into account the 24 constraints of the problem. In less than 1 second a first improved solution is obtained. In 12 minutes, the solution improves three key indicators: it increases by 11% the number of programed hours, it improves pilots scheduling balance by 6% and it reduces students schedule variability by 19%. Arturo also designed 3 new indicators and simulated taking into account a more actualized legislation which resulted in increasing by 37% the number of programed hours. He also proposed some out of the box changes in the organization that could lead to more than 70% improvement.
A word from the professor
Scheduling is one of the most common problems in Operations Research. A set of task has to be processed (in our case the classes) in a machine which is the bottleneck of the process. The whole optimization consists in getting the most out of this bottleneck. There are plenty of techniques to handle this problem. In this challenge however, identifying the bottleneck was far from obvious. Is it the instructor? Is it the planes and which ones? The reality is that in our case the bottleneck was changing depending on the solutions. The students had to analyze where the limiting constraints were and they developed techniques to dynamically identify them to make the best use of the available resources. By doing, so not only did they develop effective software but they also discovered key insights that helped the company improve their processes.

Prof. Rabie Nait-Abdallah
Leader of the Challenge
Javeriana University