Carmen and Björn Dahlberg
Jeppesen Systems, formerly Carmen Systems, is a Göteborg-based software company, specializing in resource optimization for the transportation industry. If you came here by plane, the pilot who brought you here and the flight attendant who served you were probably scheduled by Jeppesen Systems' optimizers - 75% of European aircrews are. The company's success is very much due to the quality of the optimization algorithms that are constantly improved by its research group of 20 Ph.Ds from 15 different countries. In 2006 Carmen Systems was bought by Boeing subsidiary Jeppesen, but the name Carmen lives on in the name of our crew scheduling products.
Jeppesen is a sponsor of the Dahlberg Seminar because Björn Dahlberg had a significant influence during the inception of the company's activities, which took place at Volvo Data when he was active in creating new mathematics for automotive design. Although he was never directly involved with Carmen Systems, it is very unlikely that company would exist today if it wasn't for the unique atmosphere Björn Dahlberg created in the applied mathematics group at Volvo Data in the late 80s.
I personally got involved with Volvo Data in 1986, when Björn Dahlberg gave two other students and me the assignment of implementing Karmarkar's algorithm for linear programming. Karmarkar had just released his paper where he proved polynomial convergence of linear programming with his algorithm, which he claimed was also competitive in practice. These claims created headlines and a lot of controversy in the optimization community, since AT&T Bell Labs couldn't support the claims with any evidence, but kept filing patents and threatened to go after other implementers with lawsuits. Perhaps it was because of these irritants that Björn thought it was worth spending some of Volvo's time and money on another implementation. I felt very privileged to have all that time to spend on such a hot subject.
It was so different from anything else going on at Volvo Data. Some fundamental research in mathematics was supposed to be converted into a computer program, which was a key component in another Dahlberg-inspired software, which was going to create more beautiful cars. It takes a lot of charisma to pull off such a project in a conservative industrial environment. Björn Dahlberg had that charisma. He left me with a feeling that good mathematics was dramatic and powerful and it was the only thing that could really lead engineering and technology forward in a big way.
The automobile surface design project (SLIP) created some nice surfaces but it never became very popular among design engineers. Karmarkar's algorithm however, made headlines again in the US for saving around $30 million annually on optimized crew scheduling for Delta Airlines. By coincidence our office neighbours at Volvo Data did some editing software for crew scheduling, and I soon found myself using the rhetoric I learned from Björn Dahlberg to convince Lufthansa that optimization would make a major difference for them.
Karmarkar's algorithm was outperformed by a new algorithm created by Dag Wedelin, who was also supervised by Björn for a short time. We won major contracts with Lufthansa and SAS in direct competition with AT&T, and I was probably the one gaining most from the commercial failure of SLIP. The brilliant mathematical software engineer Tony Elmroth, who made SLIP work, joined my team at Carmen Systems and turned our crew scheduling optimizer into a very successful product.
So it is coincidence that links Björn Dahlberg to Jeppesen Systems, but the spirit he once brought to Volvo Data is still very much alive at our company. The idea that advances in fundamental mathematics can make a huge difference to industry has become reality at Jeppesen Systems. It was real in the case of Dag Wedelin's algorithms and it has become real again in successors of Tony Elmroth's programs. To us the future depends on new mathematical advances and it is therefore very relevant for us to sponsor an event that has the purpose of promoting excellence in mathematics. We look forward to seeing the Dahlberg Seminar become a source of inspiration for new, brilliant, mathematicians and mathematical software developers.
Erik Andersson, Head of Product Management
Jeppesen Systems AB