Optimizing floating crane operations for efficient bulk product transshipments on inland waterways

Abstract

In the realm of global supply chains, the optimization of floating crane operations for bulk product transshipment via inland waterways emerges as a crucial necessity to address economic, operational, and environmental imperatives. This research identifies a significant gap in existing methodologies for the scheduling, routing, and assignment of floating cranes, which are essential for improving efficiency and sustainability in maritime logistics. To bridge this gap, we propose the Reinforcement Learning Variable Neighbourhood Strategy Adaptive Search (RL-VaNSAS) algorithm, a novel integration of reinforcement learning with variable neighbourhood search strategies. This advanced model aims to holistically minimize energy consumption, labor costs, and penalty costs, while simultaneously enhancing service efficiency. Through rigorous simulations, RL-VaNSAS was benchmarked against conventional methods such as Differential Evolution (DE), Genetic Algorithm (GA), and the original Variable Neighbourhood Search Adaptive Strategy (VaNSAS), revealing its superior capability in significantly reducing annual energy costs to $1,211,948, labor costs to $270,948, penalty costs to $19,948, and operational hours to 12,087. Demonstrating notable advancements in operational efficiency and cost reduction, RL-VaNSAS offers a sustainable solution to the dynamic challenges of maritime logistics, characterized by fluctuating vessel arrivals and diverse cargo requirements. The findings illuminate the critical need for innovative optimization techniques in enhancing the sustainability and efficiency of maritime logistics operations. RL-VaNSAS not only fills the identified research gap but also sets a new standard for future endeavors in global supply chain management, underlining the importance of adopting advanced optimization strategies for sustainable production and economic growth.