Imagine we have a bunch of independent agents. They will each work towards their local optimum. Under which conditions will all of these individual local optima lead to the global optimum?
For example, imagine a large set of autonomous vehicles. If each has the goal to minimize it's travel time, they will likely pursue strategies that force other vehicles to wait and this will likely cause the overall travel time of all vehicles to be longer than the minimum possible -- i.e. the global optimum.
What type of incentives/objectives could be created to ensure that the individual agents could pursue local optima that would lead to the actual global optimum?