Publication Details
Abstract
Where no less-than-truckload terminal network exists, small consignments can still share a truck if a platform assigns two to four of them to one vehicle that unloads them in sequence without rehandling. We formulate the carrier-side dispatch problem for such partial-truckload trips on a linear corridor with spur drops, en-route pickups, delivery deadlines, dynamic mass and volume limits, last-in-first-out loading, three pickup modes and a pickup time budget. Because the corridor fixes the visiting order and feasibility is hereditary, the problem is solved exactly by enumerating feasible groups and solving a set-packing model. A marginal-profit rule gives a closed-form detour radius that can filter offers. On 120 synthetic instances calibrated to the 1,250-km Tashkent–Nukus corridor in Uzbekistan, the exact method needs 40 s on average at 60 consignments. An ablation shows that trying every consignment as the trip anchor (multi-start) is what lifts greedy insertion from a 26–30% to a 4–7% mean optimality gap; ranking candidates by destination tiers makes it slightly worse. A large neighbourhood search started from anchored construction reaches a mean gap of 0.9–2.4% within a few seconds for up to 60 consignments and scales to 300 consignments. Even optimal plans leave many consignments unserved; with an unconstrained fleet 26% remain unprofitable, which supports filtering loads before they are shown to carriers.