Operations Research

Search and routing problems where the objective is expected time, not distance

Where to Look First

Not the travelling salesman's route — you stop when you find it, so the legs at the end may never be driven

When the Map Is Only a Table

The same search on a graph, where the costs need not be distances and the places need not be equally likely

How Far Ahead Is Far Enough

You must build something now. How much of the future do you have to work out first?