- Test C->Python using Python Bindings
- Create simulation program to generate traffic randomly according to a given distribution
- Starts at midnight and goes to next midnight
- Follows normal distribution
- Random variation at any given time
- 2 intersection possibility
- Pass-through intersection, where the majority of traffic on one lane and moves straight through
- Random, where traffic evenly spawns and moves in all directions
- Research current traffic algorithms
- Create main program
- Test
- Valgrind
- GUI
- Python Handler
- Generate 5 traffic files
- Runs the standard model on each one, records penalty
- Trains Neural Network on the 5 files for 5 generations
- Save penalty to output file (with epoch number)
- Restart (Do this 1,000,000 times)
- To step it up, we can increase the traffic files we do at 1 time, or do 1 traffic file for 1 generation, and do more generations overall
- Learn LaTeX to create final report