Darkness 2010-11-10 12:00:00 UTC
 
Twilight 2010-11-11 12:00:00 UTC
 
Daylight 2010-11-12 12:00:00 UTC
 
Finish 2010-11-17 12:00:00 UTC

Sailing Home - Winners

The Fall 2010 MATLAB® Programming Contest has come to an end. We hope you enjoyed it as much as we did.

Thanks to all the participants and spectators who have helped make the Sailing Home Contest such a great success. Congratulations to all our winners! We also want to give a special "high-five" to those winners who are already in our MATLAB Contest Hall of Fame, yet made the effort to participate again in this contest.

André Fioravanti
Grand Prize

I was born in São Paulo - Brazil, and currently I live in France, where I am a PhD candidate in INRIA / L2S-Supélec in the field of Control of Dynamic Systems. So it is not a surprise that I have rewritten the main problem of the contest in a more mathematical way that my professors would be interested to discuss, but all ideas failed to pass the time constraint even in the local set of problems. I was very impressed when we came out of the Darkness phase to see the quality of the results that people had already obtained with no use of data fitting, specially from Nicholas.

This was my second participation in a MATLAB Contest, where the first one was the French "Concours de la Toussaint", which happened just a few days before the beginning of this contest. It was in the middle of a long French holiday, so I could take some time to read how things went in the recent contests. As we were a lot of newcomers, we stayed almost half of the competition "respecting" each other submission, which made me lead it for a long time with a single submission. Eventually someone broke this psychological barrier and things went more like they were supposed to be. For the end game, I tried to apply the technique Hannes used to win the Sensor contest. While his submission could survive 22 minutes of re-cloning, mine sent about 5 minutes before the deadline had one parameter changed and it was beaten in the last few seconds. So I can definitely say I know how it feels! :)

I could not participate as much as I'd like in this contest. Having a thesis to write and a deadline for your scholarship makes you feel bad about doing anything else in your life. ;) But for the final minutes, I remembered Alan's coefficients for the scoring formula, but for me we should have been shooting for ~70s. For some reason, in the middle of dozens of submissions in the queue, Sebastian's code seamed to have an appealing name, and I modified it only to compromise time/result as I thought it was better. Beginner's luck, one may say.

I would like to thank all the contestants and organizers, and I hope to see you all in the next contest.

David Felguera
Saturday Leap Prize

My actual name is David Felguera, I live in Madrid, Spain, and I work for the Airborne Radar Department at Indra. I'm also pursuing the Ph.D. degree at the Microwave and Radar Group of Polytechnic University of Madrid researching in Interferometric Radar Imaging techniques. I use MATLAB every day since 2004 for Radar Signal Simulation and Radar Algorithm Design.

I've taken part in the MATLAB contest since 2005 and this is the first time I win a price. However, I've learned a lot about MATLAB programming in each of them. In this contest, I couldn't focus on the problem until Friday so I took the leader at that time and started working on it. That solver was using a tweaked version of Nick's and Bradley's algorithms and selecting the best result.

I focused in Bradley algorithm because I found it very elegant and was easier to understand for me. When debugging it I noticed that, it was better to do the operations inside the loop with the whole matrix instead of indexing the useful positions. I re-coded the loop to use constant size variables inside which resulted in better performance and also allowed me to replace all size calculations with constants. This was my first Saturday Submission which reduced 20 sec. the execution time of the previous leader. I was surprised by the increase of performance obtained with that submission so I decided to invest some of the saved time doing more iterations in the loop and getting a better result with my second submission that finally won the Saturday Leap. It's really rewarding for me seeing that many of the improvements I made with these submissions remained all over the contest and ended in the grand-price winner too.

Many thanks to the MATLAB Team and congratulations for the new contest machinery and rules that made it run smoother that ever.

Sergey Yurgenson
Sunday Push Prize & 1000 Node Challenge Prize

Sergey Yurgenson has a Ph.D. in physics from Leningrad State University (Russia). His previous wins in the MATLAB Central Contest include the Tuesday Leap and the 10000 Character Challenge Winner in Peg Solitaire (May 2007).

Currently Sergey works at Harvard Medical School. He uses MATLAB for data analysis and control of data acquisition in Neurobiology research.

I like the first couple days of the Matlab contest Daylight. The code is usually quite good by that time; however it is not overfitted to the test suite, thus making it easier to develop and test new ideas and modifications.

Main code during this contest was calculating iteratively how much fuel was needed to get to different points on the map choosing the least “consumptions” points as the start of the next iteration.

Initially, it did not take into account which point is closer to the destination. My modification introduced additional parameter, which gave small preference to points close to destination without interference with minimizing fuel consumption. Preliminary tests showed visible score improvement; however, I was not sure if it was enough to overcome then current leader in Saturday Leap challenge.

I decided to hold my submission until Sunday Push challenge. I submitted the code early on Sunday, taking lead, and spent the rest on the day making sure that nobody would slowly eat up my lead by multiple small parameter modifications. It required constantly checking submissions and participating in all “parameter hunting” initiated by other players. This strategy worked, giving me Sunday Push win.

 

Nick Howe
Twilight Prize & Darkness Prize

Nick Howe teaches computer science at Smith College in western Massachusetts. He uses MATLAB extensively to conduct research on problems in computer vision.

The Gerrymandering contest (April 2004) was his first MATLAB Central Contest. He has been participating on and off ever since.