MODEL AND HEURISTICS FOR THE VEHICLE ROUTING PROBLEM UNDER UNCERTAINTY IN INTERMEDIATE MILE WITH FULL LOADS.

This paper addresses the Vehicle Routing Problem with Full Loads and Stochastic Pickup and Delivery (SFTPDP), a variant of the problem that incorporates the uncertainty arising from order cancellation. Although the deterministic problem has been widely studied in the literature, studies that consider optimization under uncertainty in this context are scarce, characterizing the scientific gap that motivates this research. To address it, a stochastic linear programming model for the problem is proposed and the use of the RKO (Random-Key Optimizer) framework as a heuristic solution approach, which includes the construction and evaluation of decoders for the problem. The computational experiments performed with randomly generated instances demonstrate that the exact model, solved by the Gurobi solver, is unfeasible in practical applications for larger instances, while the RKO found viable solutions in all evaluated instances, with an average gap of 0.15% in relation to the known optimal solutions, evidencing the efficiency and scalability of the proposed approach.