Implementation of CTM by CUDA
The previous version builds the model directly on the cell-link level, which makes it hard to scale. To solve it, the new version will focus on the lane-intersection level. During the operations on the lanes and the intersections, the related cells and links are automatically created or modified.
Another difference with the previous version is that the access field is transferred from cell to link,
so that the model can handle more complicated phases.
The difference between this repository and the one entitled hoolheart/CTM_C_Program.git is that this repository runs the simulation on GPU instead of CPU.