composing calibrated contingencies generating definitive futures exploring precise estimations mapping precise insights predicting precise forecasts mapping the future delivering contingent insights generating precise predictions assembling definitive understanding exploring accurate estimations composing probable understanding modeling accurate wisdom composing intelligent predictions composing contingent forecasts

Question

Metaculus Help: Spread the word

If you like Metaculus, tell your friends! Share this question via Facebook, Twitter, or Reddit.

What will be the score of the best performing agent in the Flatland AI competition?

Can an RL system get the trains to run on time? The Flatland Challenge is a competition in multi-agent RL tackling the Vehicle Re-Scheduling Problem (VRSP), the problem of how to efficiently schedule transit routes while taking into account real world conditions like breakdowns and delays.

Sponsored by the Swiss Federal Railways - who manage the densest mixed railway network in the world - the goal is to create RL agents that can schedule trains in a gridworld environment which minimize the total time of all operating trains.

The first round was focused on basic scheduling - how to get multiple agents (trains) to their destination quickly. The current round has a more complex problem set, with trains: - Operating at four different speeds - Entering and leaving the environment - Randomly malfunctioning, rendering them immobile for a period of time.

The best systems will be able to plan ahead and route efficiently around malfunctions.

As of Oct 28th the best performing agent on the round two leaderboard had a normalized reward of -23.010.

What will be the normalized reward of the top performing agent be at the end of the competition?

Resolution Criteria: The outcome will be determined by the leaderboard at the end of the competition.

The normalized reward is a function of: - Mean number of agents done, in other words how many agents reached their target in time. - Mean reward is just the mean of the cumulated reward. - If multiple participants have the same number of done agents they compute a “nomralized” reward as followes: normalized_reward = cumulative_reward / (self.env._max_episode_steps +self.env.get_num_agents()

{{qctrl.predictionString()}}

Metaculus help: Predicting

Predictions are the heart of Metaculus. Predicting is how you contribute to the wisdom of the crowd, and how you earn points and build up your personal Metaculus track record.

The basics of predicting are very simple: move the slider to best match the likelihood of the outcome, and click predict. You can predict as often as you want, and you're encouraged to change your mind when new information becomes available.

The displayed score is split into current points and total points. Current points show how much your prediction is worth now, whereas total points show the combined worth of all of your predictions over the lifetime of the question. The scoring details are available on the FAQ.

Note: this question resolved before its original close time. All of your predictions came after the resolution, so you did not gain (or lose) any points for it.

Note: this question resolved before its original close time. You earned points up until the question resolution, but not afterwards.

This question is not yet open for predictions.

Thanks for predicting!

Your prediction has been recorded anonymously.

Want to track your predictions, earn points, and hone your forecasting skills? Create an account today!

Track your predictions
Continue exploring the site

Community Stats

Metaculus help: Community Stats

Use the community stats to get a better sense of the community consensus (or lack thereof) for this question. Sometimes people have wildly different ideas about the likely outcomes, and sometimes people are in close agreement. There are even times when the community seems very certain of uncertainty, like when everyone agrees that event is only 50% likely to happen.

When you make a prediction, check the community stats to see where you land. If your prediction is an outlier, might there be something you're overlooking that others have seen? Or do you have special insight that others are lacking? Either way, it might be a good idea to join the discussion in the comments.