Our simulation is controlled by a discrete timer. The granularity of the timer units determines the

granularity of the road surface and the atomic car movement increments. Each car occupies two

square units of the road surface and can move one unit forward during one timer cycle. These

square units $($places$)$ form the basis of our simulation.

Place Class

This class is at the heart of the entire program. Each place represents one patch of asphalt on the

road. It contains four pointers to other places adjacent to this one. The pointers contain valid links

only if travel in that direction is permitted. So$,$ all places in the northbound lane $($except for the

last one$)$ will have links to their neighbors north of them, but no other links. When we start building

intersections, the places where two directions of travel cross will have links to both neighbors. A

place can be blocked if it is occupied by a car or if the light ahead is not green.

To display a place $($e$.$g$.,$ after a vehicle moved away$)$ we also need to record its physical location

in the display. Because the whole display can be treated as a grid specifying the row and column

information is sufficient.