Frictionless Bananas

Team name: Frictionless Bananas
Team members: Jeremy Sawicki (jeremy at sawicki dot us)
Language: C++
Source code: source.zip

The Frictionless Bananas had only one team member this year. This page (briefly) describes my 2025 ICFP Programming Contest entry.

Lighting round

My approach was to make a single exploration consisting of random moves, and then try to find a map that is consistent with the observations. To find the map, I use simulated annealing. The simulated annealing process determines two things: the map connections, and which room we are in at each time step. All problems were solved with the minimal score of 2.

Full round

Due to a combination of lack of time and lack of ideas, I did not make much progress on the full round problems.

Source code

Some notable source files:

• lightning1.cpp: The code that actually solved the lighting round problems.
• lightning2.cpp: A faster version of the above that incrementally updates the simulated annealing score, also written in the first 24 hours but after solving the problems.
• lightning3.cpp: Further tweaks after the first 24 hours.
• anneal.cpp: Some work in progress on the full round problems.
• sat.cpp: At attempt to replace simulated annealing with a SAT solver, which proved to be too slow on larger problems.
• manual.cpp: A tool that I used to manually solve aleph to get a better understanding of the structure of the full round problems.