Logistics problems like vehicle routing, container stacking, goods packing, scheduling tend to become extraordinarily difficult to optimize. For example, how can we package and transport millions of items using thousands of vehicles, while managing complex schedules, varying delivery guarantees, transport regulations, vagaries of traffic and weather? And not lose money?
We need to take literally millions of individual decisions. Just deciding which of the million items go into which of the thousand vehicles itself creates a search space of size 1,0001,000,000. Incredibly large!
Such gargantuan, “NP-hard”, combinatorial optimization problems are all too common in operations research. No known computer system can traverse the entire search space to find the minimum, i.e. the most optimal solution. However one can find minimae, and estimate closeness to the optimal.
Using advanced mathematical techniques, we have developed various logistics planning algorithms for our customers. The algorithms give very good minimae, along with details regarding closeness to optimum, in a short computational duration. Examples include packaging optimization, vehicle scheduling, route planning, meeting service guarantees, and so forth.
