ORTEC Wins DIMACS Challenge on vehicle routing with time windows

Brilliant ideas and powerful engineering tricks

As ORTEC, we participated in this challenge to demonstrate our vehicle routing optimization expertise in a fair competition. “And ultimately, we won because our existing solutions and our colleagues already make us world champions in our field,” says Leendert Kok, Chief Science Architect at ORTEC. “The algorithm built for this challenge consists of state-of-the-art methods, brilliant ideas inspired by our production algorithm, and powerful engineering tricks that we apply across our products.”

We’re looking into whether and how we can apply our ideas to customer cases,” says Wouter Kool, OR Engineer and team leader, “and the goal is of course to bring these learnings back to our products. But we’re also publishing our project as an open-source solver, to encourage academic research on vehicle routing.”

Built in one week

As one of the challenge’s organizers also mentioned on Twitter, what is perhaps most impressive about the team’s achievement is that most of their specialized solver was built during one week of hackathon. "The ORTEC expertise and passion shines through these results,” says Wouter Kool. “Our achievement shows what ORTECers are capable of when it comes to optimization."

As a true ORTECer, Wouter describes himself as someone who likes a challenge and has a passion for pushing the limit and getting the best results. “The point is not winning per se, but stimulating new ideas. And I believe in the importance of keeping up with Academia – state-of-the-art methodologies guarantee the best performance in our products.”

Fueling lines of research

DIMACS is a world-leading platform for catalyzing and conducting research and education in mathematics, modeling and algorithms. Their Implementation Challenges aim to improve the practical performance of algorithms for “important problems […] that are hard in the theoretical sense." The puzzles proposed in such challenges are somewhat removed from the constraints of the real world but prove their value in discovering solutions to complex problems.“The DIMACS Implementation Challenges and other contests we participate in, or sponsor bring complex problems into the limelight,” adds Joaquim Gromicho, Science & Education Officer at ORTEC. “Good algorithms are robust against the differences of real-life business cases, and give good outputs even when we modify the constraints. Participating in these challenges fuels lines of research with a sizeable, tangible impact on the world.”