For several years, I have organized the SpotSeven Lab Optimization Challenge for students of the “Optimization” and “Numerical Methods” lectures in the “Automation & IT” master’s program at TH Köln. This year’s challenge presents a complex ten-dimensional optimization problem. Students can send unlimited requests to an evaluation server to receive a response for their input vectors. However, there are two catches:
1. Noise: The server returns a noisy solution and
2. Instability: A server crash is simulated, meaning that requests fail with a certain probability.
The students’ algorithms must be robust enough to handle these uncertainties while searching for the optimal solution. The top three teams will receive a certificate of achievement.
Leaderboard
We maintain a leaderboard to track the best student results, which is updated weekly. Submissions are due every Tuesday at 9:00 a.m.
Deadline 1: December 23, 2025
Deadline 2: December 30, 2025
Deadline 3: January 6, 2026
Final Deadline: January 13, 2026
Since the true global minimum of the function is unknown, it is particularly exciting to follow the trajectory of the results. After the challenge concludes, we will analyze the different algorithmic approaches the teams developed.
Current Standings
Week 1: The winner of the first week is Team “EigenSquad” with an objective function value of 10.51045684. Note: This is the non-noisy value I verified based on their submission.
Week 2 (Ongoing): We are currently in the second week of the challenge, and the previous best has already been surpassed! The new leading value, achieved by Team “Optteam”, stands at 9.84249.
Good luck to all teams for the remaining weeks!