Vinti, Cinzia (2010) The crossdocking distribution strategy: a mixed integer linear programming model and its extension for a distribution process through container terminal. [Tesi di dottorato] (Unpublished)
Visibile a [TBR] Repository staff only
|Item Type:||Tesi di dottorato|
|Uncontrolled Keywords:||crossdocking, distribution strategy, mixed integer linear programming model, container terminal|
|Date Deposited:||22 Mar 2011 08:53|
|Last Modified:||30 Apr 2014 19:44|
Abstract Crossdocking is a logistic strategy used to improve the effectiveness of goods distribution by aiming to decrease inventory and transportation costs. A distribution network is an integrated set of suppliers, distribution platforms and customers where strategic, tactical and operating decisions related to a single player could produce effects on some (or many) others. The state of the art on crossdocking logistic strategy highlights a mismatch between the description of a crossdocking platform and its mathematical formulation. Unloading, loading, sorting, consolidating, storing, labelling and handling are the activities performed within a crossdocking platform which are not included in the modelling part. Generally the authors include just the storing activity. These activities could be relevant for some problems such as scheduling, layout and distribution whereas they could be considered irrelevant for some other problems which do not reach the detail level on the internal functioning of the platform such as the location problems. Among these problems, the PhD dissertation deals with the distribution flow problem which consists in determining how to send products from suppliers to customers through crossdocking platforms. The activities performed at the platforms are associated to costs and capacity constraints for the available resources. For this reason a crossdocking platform cannot be represented by a single transshipment node. In order to take into account these features, the crossdocking platform is modelled with a transshipment nodes network: a receiving node which stands for the activities performed on the incoming products, a storing node, which stands for the inventory activity and a shipping node standing for the activities performed on the outgoing products. The literature review on distribution flow problems for the crossdocking strategy underlines another mismatch: the lack of a specific constraint for the transportation efficiency. A model is proposed to fill these gaps. The main idea is that the crossdocking, with appropriate differences, can be formulated as a Fixed Charge Network Flow Problem, well-known NP-hard problem. Two exact approaches, a Branch and Bound and a Branch and Cut algorithm, have been developed in order to solve the problem. These two procedures are customized on the network features and they are different from the default procedures embedded within the most diffused optimization software: Xpress and Cplex. Some pseudo-random instances have been generated with the aim to test the developed model and procedures. For the large instances, the obtained results have been compared with those obtained by the optimization software. The obtained results demonstrate the effectiveness of the developed procedures as well as of the crossdocking strategy in terms of average level of inventory. The second part of this PhD thesis deals with an application: the case of the distribution process through container terminal. A new model is proposed, extension of the previous one, which allows to take into account the specific features associated a container terminal. Once again, the model is formulated, with appropriate differences, as a Fixed Charge Network Flow Problem. In this case the model is validated with a real instance extracted by the current functioning of the container terminal of Naples (Italy), which represents the case study. Like the previous model, the obtained results validate the crossdocking strategy for the management of the terminal yard. This strategy, in fact, allows to drastically cut the average inventory time.
Actions (login required)