Minimum Integer Weight Calculator

NOTE: this page was mirrored and edited from MIT's page because that version seemed to be corrupted. Program contents are copyright MIT.

How to install this program:

1. Download the weights1-01.jnlp file to your computer (Right-click and choose "Save Link to Disk" or "Save Target As"). Updated 09-30-2005 for Java 1.5.

2. Locate the file and double-click it to run.

(You must have Sun Java installed. Obtain from http://java.com/en/download/manual.jsp if you don't already have it. The program specifically uses Java Web Start)

 

Problems I've seen

I'm getting weird Baron-Ferejohn weights: they look wrong. (Updated: 9-30-05)
Baron-Ferejohn weights are calculated given that each party has an equal chance of being picked as formateur. This assumption may produce counter-intuitive results. We are currenly investigating the possibility that these weird results are just plain wrong. I'm also working to give users a choice of formateur probability functions. Keep looking here for further updates. Until then, consider B-F results suspect, play around with games like (6,5,4,3,2,2) and let me know whether you think the program is right or wrong.

I have an older version of Java, such as 1.4.2, and your program doesn't work.
Here is the older version. Alternatively, use this link to download the newest version of Java.

I have a problem with starting the program that you don't describe here
Check out this help page on Java Web Start.

The program freezes on large games, when I check to logs I see a memory error.
You ran out of memory because your game was too big. The algorithm (most likely) did not break, it just needed more RAM. I have tried in vain to fix this problem previously. Some time in the future, I'll try fixing this on the most recent version of Java Web Start. Thus far, this is the only known bug for minimum integer weights.

How to use the program

I realize the program's GUI isn't the best. Here's an attempt at a better step-through.

The rest is just output.

*Since in my paper I showed that minimum integer weights are not unique, it is conceivable that two algorithms will arrive at different answers. This outcome will occur rarely. If you think the program gave you an incorrect answer, please let me know: aaronbs@alum.mit.edu.