THESIS
2006
xiii, 102 leaves : ill. ; 30 cm
Abstract
Crossdocking has been catching on for the past several years as more and more third-party warehouses and distribution centers add it to their list of services since crossdocking with its many advantages, including cutting costs and trans-portation time, is expected to capture their undivided attention in short order. The application of crossdocking functions shifts focus from maintaining inven-tory storage to flow-through inventories. Crossdocking is an important concept because it can greatly decrease inventory storage and reduce the product flow line between the manufacturer, the warehouse and your customers. In a global level, how to schedule the transhipment, how to allocate the cargoes to the trucks in LTL (less-than-truckload) transportation segment, and how those crossdocks coope...[
Read more ]
Crossdocking has been catching on for the past several years as more and more third-party warehouses and distribution centers add it to their list of services since crossdocking with its many advantages, including cutting costs and trans-portation time, is expected to capture their undivided attention in short order. The application of crossdocking functions shifts focus from maintaining inven-tory storage to flow-through inventories. Crossdocking is an important concept because it can greatly decrease inventory storage and reduce the product flow line between the manufacturer, the warehouse and your customers. In a global level, how to schedule the transhipment, how to allocate the cargoes to the trucks in LTL (less-than-truckload) transportation segment, and how those crossdocks cooperate well with each other to achieve the global optimal performance in a complex crossdocking distribution system are all very important issues in daily operations. On the other hand, for a local level, how to assign the docks to the trucks is another key issue in crossdocking since good assignment can maximize the throughput of the crossdocking operations and minimize the total operation cost. This thesis is motivated to model, analyze, and provide algorithms on solving above two sets of problems. In particular, a major challenge in making supply meet demand is to coordinate transshipments across the supply chain to reduce costs and increase service levels in the face of demand fluctuations, short lead times, warehouse limitations and transportation and inventory costs. And transshipment through crossdocks, where just-in-time objectives prevail, requires precise scheduling between suppliers, crossdocks and customers. In this work, we study the transshipment problem with supplier and customer time windows where flow is constrained by transportation schedules and warehouse capacities. Transportation is provided by fixed or flexible schedules and lot-sizing is dealt with through multiple shipments. We develop polynomial-time algorithms or, otherwise, provide the complexity of the problems studied. Another two kinds of transshipment problems in LTL transportation segment, which are formulated as Integer Programming (IP) problems and proved to be strongly NP-complete, are also studied in this thesis. We develop a two-stage heuristic algorithm to obtain good solutions efficiently. For local level, we mainly study truck-dock assignment problems with two versions. The first version is to consider the over-constrained truck dock assignment problem with time windows and capacity constraint in crossdocks where the number of trucks exceed the number of docks available and the capacity of the crossdock is limited, and where the objectives are to minimize the total shipping distances. However, in the second version, the big difference is to take the operational time for cargo shipment among the docks into considera-tion, so that the problem feasibility is affected by three factors: the arrival and departure time window of each truck, the operational time for cargo shipment among the docks, and the total capacity available to the crossdock. Both of them are then formulated as IP models. We find that as the problem size grows, the IP model size quickly expands to an extent that the ILOG CPLEX Solver can hardly manage. Therefore, two meta-heuristics approaches, Tabu Search(TS) and Ge-netic algorithm(GA) , are proposed. Computational experiments are conducted, showing that meta-heuristics, especially the Tabu Search, dominate the CPLEX Solver in nearly all test cases adapted from industrial applications.
Post a Comment