An improved freight transportation planning system for less-than-truckload operations of a third-party logistics carrier

Abstract

Less-than-truckload (LTL) transportation is a widely used shipping modality within the logistics industry, facilitating the movement of loads insufficient in size to occupy the entirety of a truck's capacity. Drawing on the real-life problem of a third-party logistics (3PL) carrier, which navigates an expansive transportation network and unpredictable demand patterns, we propose a centralized planning approach for LTL transportation, taking advantage of consolidation opportunities at cross-dock hubs to minimize long-term freight costs while ensuring on-time delivery. We present an integer linear programming (ILP) formulation to generate a multi-day plan with known demands and a daily planning problem (DPP) model to be solved daily, emulating the available order information in a real planning scenario. DPP optimizes truck routes and loading plans, incorporating a penalty cost for postponed transportation requests. We devise a novel multi-stage matheuristic integrating geographical decomposition with tractable integer programming models and column generation-based heuristics. With this approach, the DPP is solved in fewer than 2 h of run-time for real problem instances with as high as 1400 transportation orders. We benchmark our results against a baseline heuristic that mirrors the firm's existing planning approach through a multi-period simulation. Our comparative evaluations reveal a noteworthy 7.8% improvement in long-term freight costs, concurrently maintaining adherence to the 98% on-time delivery target.