Workforce routing
April-May 2021
About the company
Baobab Soluciones is a company specialized in Operations Research and Data Science and operates in a variety of sectors like logistics, planning and human resources. It's a spin-off from Universidad Politécnica de Madrid and counts as clients some major European multinational corporations like Airbus, Alstrom, Repsol, Iberia, etc.
The Challenge
Baobab's challenge comes from one of its presale prospects, a major European telecommunication operator. The company must audit its stores every three months. The problem consists in optimizing auditors’ routes to visit more than 2000 stores in Spain. All stores must be visited within a limited period of time and the objective is to minimize costs. What makes the problem complex is that one must determine the route of each auditor, while taking into account some very specific considerations like regional holidays, human resources regulations and also decide where the auditors will spend the night at the end of each day: either go back home or stay in a hotel to avoid additional traveling costs. Baobab realized a quick preliminary proof of concept to solve the problem knowing that there was still room to improve. The Challenge with Javeriana was an opportunity to explore to which extent the problem can be further optimized.
45%
12%
decrease in
overall distances
makespan reduction
8%
of overall cost reduction
500x
faster execution time
Winners of the Challenge
Tania Lorena Granados
Sandra Diaz
The software Tania and Sandra developed reduces by 24% the number of required auditors in just 4 seconds of running time. The approach is based on finding an equilibrium between reducing the working days and minimizing the overall distances. Their solution results in saving 16 working days and 10% of the overall distances. In addition to their software, Tania and Sandra analyzed Baobab approach and proposed two modifications which result in around 6% reduction in costs and 25% in the number of required auditors.
Their work has been submitted to an international operations research conference in Spain.
Maria Alejandra Prieto
Daniela Catalina Jimenez
Maria Alejandra and Daniela's software manages to reduce the overall costs by 8%. Their approach is based on a first phase of clustering that prioritizes the opportunity cost of inter cities distances. The solution results in 15% less auditors, -10% overall distance and -6% makespan. All Baobab instances have been improved and the execution time is 500 times faster compared with Baobab's original approach.
Their work has been submitted to an international operations research conference in Spain.
Andrea Amaya
Juliana Cardona
En just a few seconds of running time, Adrea and Juliana's software reduces the total cost by 6%. In more details, their solution reduces the number of required auditors by 45% and improves significantly (85%) the match with auditors preferences. They also discovered two key insights to reduce costs, identifying the relation between holidays, hired auditors and the number of cases where half workdays are required.
Catalina Bohorquez
En less than 15 seconds, Catalinas's software reaches a solution with a cost 11% lower than Baobab's reference solution. The approach is predicated on the idea of efficiently creating groups of municipalities to optimize distances and use more effectively the auditors. This approach reduces overall distances by 32% and reduces the number of required auditors by 21%.
A word from the professor
This was a very interesting Challenge from so many perspective. From the pedagogical perspective, it was very instructive for my students to see how Operations Research professionals approach and solve real problems of the industry. It was also the first time that the Challenge got international and we were very pleased to work with a European company. Finally, on a technical perspective, the problem in and of itself is a VRP (Vehicle Routing Problem) which is a classic problem in optimization. However, all the specific constraints made the usual techniques not very effective. And the students had to dig deeper in the core nature of the problem using some creative and even counter intuitive approaches to improve Baobab's solution.
Prof. Rabie Nait-Abdallah
Leader of the Challenge
Javeriana University