Kaggle’s Traveling Santa Optimization Challenge

Print Friendly, PDF & Email

???????????????????????????????????????????????Kaggle, the machine learning challenge organization, is getting with the spirit of the season by launching a new competition: A non-denominational Seasonal Convex Optimization Problem, v.2.01. All data scientists are welcome to participate. The competition will run from Dec. 2, 2013 until Jan. 26, 2014. There are three prizes, two for $4,000 and one for $2,000. The sponsor is MathWorks, creators of MATLAB® and Simulink®.

Problem Description

Given a list of presents, pack them in Santa’s sleigh as compactly as possible and in the best order possible.

The sleigh and presents are discretized and described in units of the fundamental length unit ℓ. The sleigh is 1000 x 1000 with infinite vertical extent as needed by your highest placed present. The cells of the sleigh go from 1 to 1000 for the length and width, and 1 to infinity in height.

Presents come in random sizes and are represented by their extent in the x, y, and z dimensions. Each present has a PresentId, a number which determines the order in which it is to be delivered. Ideally, Present 1 is to be delivered first, Present 2 second, etc.

Present 1 is 2 x 5 x 3 ℓ^3 and Present 2 is 243 x 207 x 73 ℓ^3. Presents can be packed in any orientation provided they are parallel and perpendicular to the x-y-z axes, meaning they can be rotated in any direction by multiples of 90∘ but not, for example, by 60∘.

This sounds like a fun way to spend time during the holiday season, so sign up now, download the data, design your algorithms and start making submissions.

 

Speak Your Mind

*