FreeSeat is a Common Lisp application designed to aid in planning seating for events. If people were reasonable and got along, this would be a snap and this application would be wholely unneccesary. People, however, are not reasonable and do not get along. Some people have "spouses" or "significant others" with whom necessarily must sit, and some people have "rivals" or "enemies" with whom they necessarily cannot sit. Similarly, some people have "friends" with whom they want to sit, or people who "annoy them" with whom they do not want to sit. There are varying degrees of desire in these cases.
Rather than encourage everyone to work out their differences peacefully and make friends with one another, why not use the power of mathematics to construct socially advantageous seating arrangements? While a provably optimal solution would be incredibly difficult to find, when it comes to dinner parties, "close enough" is close enough. This program ensures that as long as all groups of people who must sit together are together and no two people who cannot sit together are together, and just does its best with everyone else.
It does so via an optimization technique known as a Markov chain Monte Carlo method. Considering the constraints of the problem, it explores the state space in search of the seating arrangement with the greatest utility.
The program takes input in the form of a CSV file (see the FreeSeat launch party document for an example) that has a field for a guest's name, a list of people with whom they must sit, a list of people with whom they cannot sit, a list of people that they should sit with (scored 0-99), and a list of people that they should not sit with (scored 0-99). It groups them currently at tables of size 5, but this is trivially changable. The program is invoked using "clisp load.lisp". Replace clisp with your favorite Common Lisp interpreter, but it's only been tested in clisp.
This program was written for a hackathon, so please ignore the hacks used to get it working in a short timeframe (i/o capacities in particular) and bugs that escaped our rigorous and extensive (read: near-nonexistent) testing.