Matheuristic algorithm for dock planning problems considering tandem and parallel methods in the shipbuilding industry

Abstract

Docks are crucial production resources that significantly impact shipyard productivity. This study addresses a real-world dock planning problem (DPP) focused on ship production scheduling within docks. Unlike traditional scheduling issues, the DPP involves managing multiple concurrent tasks due to the use of tandem and parallel methods common in the shipbuilding industry. This distinctive feature introduces unique considerations and opportunities for optimizing production processes within shipyards. Despite its importance, research on DPPs is limited, with dock planning typically depending on the expertise of planners. This study formally defines a DPP as a mixed-integer programming (MIP) model, incorporating tandem and parallel methods. It also introduces MathLNS, a matheuristic algorithm that combines MIP-based large neighborhood search with heuristic methods, as an efficient solution approach to this problem. The effectiveness of the proposed model and algorithm is evaluated using real-world data from one of the largest shipbuilders globally. The results demonstrate significant improvements over the planner-driven approach. Notably, the developed algorithm has been integrated into the company’s production planning system, resulting in improved dock planning efficiency. Furthermore, the proposed approach offers potential benefits for other shipbuilding companies, enhancing operational efficiency in dock planning.