Operational decision models for maritime routing and scheduling of bulk products using mathematical programming and simulation

Logistics constitute an important activity for organisations dealing with physical goods. The majority of research in transport logistics however, has been in land-based transportation. In comparison, maritime logistics has received little attention even though ship fleets require greater capital investments and operating expenditures. Futhermore, the overwhelming majority of goods by volume, is carried by sea. In maritime logistics, container shipping has received the greater part of the research attention while the shipping of bulk products such as grain, ore and liquids has remained relatively unexplored. This research explores two important aspects of bulk shipping. The first part involves the development of a model for the determination of the optimal routing, schedule and loading plan for a fleet of ships carrying bulk products. A mixed-integer linear programming formulation is developed to determine routes, departure-arrival times and quantities to be loaded or unloaded for a fleet of heterogeneous ships carrying multiple products between harbours. Some harbours are production points for particular products while others are consumption points. Essentially the problem is to determine the route, timing and quantities that minimise the transportation cost such that the inventory levels at harbours, affected by consumption and production, are maintained within storage capacities. The problem is complicated by multiple-berth harbours which require the formulation to be extended so that inventory levels are correctly tracked when multiple ships concurrently load or unload products at a harbour. Secondly, the developed formulation allows flexible use of onboard compartments, removing the need to dedicate compartments to specific products. Third, an extension to the formulation allows a harbour to serve as a redistribution point to other harbours even though it is primarily a consumption point. This makes it possible to send large loads to a redistribution harbour and then send smaller loads to nearby harbours. Fourth, an extension is developed to allow individual loading and unloading rates to be specified for each ship–harbour pair. Formulations in literature assume a constant rate but real-world rates are determined by the performance of onshore and onboard loading equipment. The formulation seeks to minimise the total transport operating cost within the planning horizon including fuel and port charges, loading /unloading costs and other fees. In addition, the formulation is able to take into account costs associated with ships that are chartered instead of owned by the operator. Variations of the formulation are used in solving nine different configurations. Results demonstrate that the routing, timing and inventories are accurate. Further, the flexible-use compartment and redistribution harbour extensions are shown to provide substantial improvements in transportation costs and lead to better utilisation of onboard compartments. The formulation provides optimal decisions for integrated routing and inventory operations, leading to better use of transport assets and storage facilities. While the first part of the research looks at the overall planning of the distribution system, the second part focuses specifically on the ship queuing problem just prior to berthing and. This is motivated by the long ship queue and waiting time, leading to demurrage costs in the millions of dollars experienced at several coal export terminals in Australia. The purpose of this research is to explore alternative priority rules for routing ships to the berths from a queue outside the terminals. At present, ships are routed by turn of arrival. A discrete-event model is developed to represent a section of the coal supply chain leading up to the loading operation at the terminals as well as ship arrival and queuing. The model takes into account stochastic features of the operation as well as maintenance, breakdown and delays. A case study for two coal export terminals is presented and the performance of alternative priority rules at varying demand requirements is explored. Results indicate that any of the alternative priority rules tested performs substantially better than the existing priority rule in terms of the waiting time profile, ability to absorb higher throughput requirements and berth utilisation. These alternative rules provide a promising solution to an especially costly problem.