Re-uploaded. Should be there now.

Almost nothing that can be learned from ants will be useful in avoiding traffic jams with cars. The reason is that ants, like pedestrians, can walk along almost any path along a surface. If there is a blockage along a path, they can go around it and try to find a new path. Cars cannot do anything similar because they are largely constrained to a unidirectional 1D path, especially on highways.

That is why highway traffic can be modeled quite well by physicists: it is essentially a 1D model where each agent can be described by very few parameters. The jam/no jam scenarios can fluctuate in time, and often don’t have a static cause (such as an accident). Instead, they are often driven by near collisions, and the backward propagating waves that ensue.

Relying on the decentralized individual to solve this program is way too idealistic. We must have a centrally organised system. All we have to do is use GPS systems in cars and then have a centralised system control the traffic flow by physically blocking off overly congested roads and providing optimised routes to the drivers of the vehicles. Since it is a non-linear system, if we have such a system in place for 50-60% of the vehicles, it would already make a huge difference.