Supplying of retail outlets is a process driven mainly by commercial criteria (e.g., just-in-time deliveries, warehouse outsourcing), which often disregards energy, environmental and social impacts produced by transport of goods. In urban areas, significant efficiency and environmental improvements could be obtained by introducing operational centres (called city terminals or urban logistic platforms), located in strategic points, where goods coming from outside are collected, consolidated and delivered to the city centre by means of small and clean vehicles optimally charged and operated by applying ITS principles (Thomson, 2003). However, an urban distribution centre introduces a further breakdown into the delivery and transport process, which could produce longer delivery times and increase distribution costs, clashing with the current trend of providing just-in-time supplies. As a consequence, conditio sine qua non for the success of urban logistic centres is that their operations were so efficient and quick that re-combination of loads does not affect the overall quality of service required by the market. Achieving this goal requires optimising the functional lay-out of the logistic centre and extensive application of information and communication technologies, like tracking and tracing of vehicles and parcels, individual route guidance, Electronic Data Interchange protocols, real time optimisation of routing and scheduling. Since goods can arrive atthe logistic centre even without previous notice, it is also necessary that all operations will be coordinated and optimised in real-time in order to promptly react to unexpected events. The proposed paper introduces two software tools carried out jointly by ENEA, the Italian Agency for Energy,New Technologies and Environment, and the Department of Hydraulics, Transportation and Roads at the University of Rome "La Sapienza". From a mathematical point of view, both problems are classified as NP-hard combinatorial and require heuristic procedures for a solution. Specifically, the operational phase is very complex, as the process continuously evolves during the day and requires taking decisions in a very short time; in fact, the operational scheme considered here assumes that parcels arrive randomly at the city terminal and have to be delivered within the daily time windows required by the final consignee. Both software tools are based on a mathematical framework where the decision process is formulated as a multi-criteria problem and values of design variables are computed by solving location and routing problems on a graph representing the whole road network. The solution procedures combines deterministic (Analytic Hierarchical Process, clustering techniques) and stochastic (multi-string genetic) algorithms. The software tools have been tested in simulation on two Italian towns, Terni and Cosenza, whose Administrations are taking into consideration innovative solutions for freight transport and the results have been very encouraging, as up to 55% reductions in the total distance covered have been obtained. For the covering abstract see ITRD E135582.
Abstract