Last-mile delivery is a complex business with low margins. Be it food, grocery distribution to doorsteps or parcel/courier delivery from hubs to doorsteps or daily replenishment of supermarkets from warehouses. These operations become more complicated if we consider time windows for delivery, truck size restrictions in some neighborhoods, driver shifts and break regulations, SLA, fixed and variable costs of trucks, etc. In mathematical terms, the above problem is NP-Hard. Saddle Point implements a cloud-based solution which is developed using Meta Heuristics (Simulated Annealing and Tabu Search) and Local Search techniques. This solution produces high-quality solutions compared with simple heuristics like Clarke and Wright and is very scalable compared to models developed in mixed-integer programming.
