PhD Candidate, Industrial Engineering and Operations Research · Penn State
A broker in a truckload market sees loads arrive over time and has to decide which carrier serves each one, when it is scheduled, and what it is priced at. The decisions are not independent. Committing a carrier now changes what capacity is left when the next load appears, so the problem is sequential rather than a sequence of separate assignments.
I treat it as an approximate dynamic program. The market is cleared exactly at each decision point, and a value function approximation carries forward what remaining capacity is worth, learned from the duals the clearing problem itself produces. Hours-of-service rules and multi-load chains make the scheduling subproblem combinatorial in its own right, which is where most of the algorithmic work sits.
Papers from this line are in preparation.