Serious hydraulic infrastructure management game
Do you want to help a distribution company decide on its preventive renovation policy to avoid water leaks using Machine Learning?
The starting point of the game is the failure history of a water network (2017-2021) with physical data (material, diameter, length) and other data (age, failure history, and yearly failures) for each pipe and year.
The objective is to recommend which pipes to replace (given a budget) in order to minimize failures the following year (2022). This is a real and relevant problem (FLUENT Project) on which the research group cs2ac has been working in recent years. It is a problem aligned with the research line of the UPC-TAIGUA chair, as it connects urban hydraulics, digitalization, and hydroeconomics.
- On Tuesday, May 13, from 14:00 to 15:00 in classroom 3.2 of TR5, as part of the cultural week of the Terrassa Campus, the Predictive Model Simulation Game will be presented.
- Registration is required through the form.
- If you manage to reduce the number of leaks the most for the following year with a fixed budget, you will win a great technological prize.
- Based on real data and with access to ML models (Logistic Regression, Random Forest, and Neural Networks) you will have to make decisions to train the models and apply their results according to the established limitations.
- Bring your computer, nothing needs to be installed as the game runs on our server.
- Whether you win or not, you will certainly learn.
The winner of the cultural week contest at Terrassa Campus was Elies Grancha Giménez
 |
 |
This game was created using algorithms and methodologies
developed in the FLUENT project. https://cs2ac.upc.edu/ca/projectes-de-recerca/accio_pagines/fluent2.0
Currently, to play online the UPClink VPN must be active. We are working to provide universal access. The game is available at: http://esaiict.upc.edu:8080
Share: