Heat home Save from climate

• GIF

  four room heat

• 

  learning curve

• GIF

  multiroom

*Inspiration *

As some part of world face issue of climate change, I want to develop an application of AI that use reinforcement learning that could help household in Vancouver reduce their carbon footprints. However, high amount of energy come from hydropower. Even though, household energy predominantly relies on fossil fuel. making it problem for world to tackle.

What it does

Simple 2D heat transfer model to stimulate house in winter, and trained reinforcement learning agent to optimize heating schedules so as to reduce waste and maximize the comfort of stimulated occupants with their own schedule and temperature preference.

How we built it

First try to trained heat data using reinforcement learning that optimize heating schedule.

Challenges we ran into

Heat transfer modelling, reward shaping

Accomplishments that we're proud of

gym platform that I am not familiar with it and I try to implement it on my own

What we learned

learn neural network and reinforcement learning algorithm

What's next for Heat home Save from climate

Investigate use of reinforcement learning to encourage human to heating their home less 1% saving per degree.

Built With

• gymflow
• numpy
• python
• ray