Design and project appraisal of container terminals may be carried out through two main approaches. optimization or simulation. Simulation can help to achieve different aims: overcome mathematical limitations of optimization approaches, support and let computer generated strategies/policies more understandable, support decision makers in daily decision processes through a “what if” approach. Although in literature numerous efforts may be found in the field of simulation of a container terminal, most of the existing papers do not pay great attention on the model set-up, its calibration and its validation, but they are only focused on the application and/or on the comparison of design scenarios. In a container terminal may be distinguished macro-operations, operations and activities. Macro-operations are set up by operations; operations are set up by handling activities. In such a classification, dependent on the level of aggregation chosen, a performance function can be associated to an entire operation (e.g. vessel loading) or to a single handling mean (e.g. a quay crane that load a vessel). In all the cited cases, performance functions can be deterministic or stochastic and are highly influenced by containers typology (e.g. 20 feet vs 40 feet) and by their state (full or empty). Although the estimation of performance functions should be one of the main issues of all container terminal applications, such a problem does not seem to be treated deeply in most of the application existing in literature. If on the one hand, many contributions do not present any information on performance functions used, the remaining contributions carry out very simple approaches (deterministic) and/or give scanty information: on the estimation approach pursued, on experimental data used, on parameters estimated and on parameters value. In this paper, drawing on the model architecture proposed in a previous contribution by the same authors, a discrete event simulation model is proposed for the Salerno container terminal (one of the major private container terminal in the South of Italy) in order to address some of the topics introduced before and, in particular, to cope with the following issues: 1. calibration of stochastic distribution function to estimate terminal handling activities, 2. comparison of different stochastic estimation approaches ( Moment, Maximum Likelihood and Kolmogorov-Smirnov), 3. analysis of the effects of different estimation approaches (deterministic, stochastic aggregated, stochastic disaggregated) on the estimation of the whole terminal performances. The main container terminal operations have been subdivided into elementary activities and each activity has been analyzed individually. In particular, the handling means considered have been: quay cranes, yard crane and reach stackers. The activities taken into account have been:  as regard quay cranes: loading time from dock to vessel, loading time from shuttle to vessel, unloading time from vessel to dock, unloading time from vessel to shuttle;  as regard yard cranes: unloading time (to shuttle/truck), loading time (from shuttle/truck), unloading time (to stack), loading time (from stack), trolley speed (with container), free trolley speed, crane speed;  as regard reach stackers: unloading time from shuttle/truck, loading time to shuttle/truck, stacking time (to tier). All of them have been analyzed for the following container typologies: undifferentiated, 20’, 40’, 2 x 20’, full and empty. Starting from the empirical frequency distributions, different distribution functions have been hypothesised (Gamma, Normal, Weibull, LogNormal) and the corresponding parameters have been calibrated starting from real observations obtained from SCT (more than 1.000 vessels monitored) and from an integrative survey (more than 3.000 containers monitored). The calibration stage has been carried out through the implementation and the comparison of Moment, Maximum Likelihood and Kolmogorov-Smirnov estimation methodologies. For each handling means and for each container typology, deterministic values and distribution function have been obtained, have been compared and have been implemented in a discrete event simulation model. The model simulates operations of SCT, and it has been used to simulate different modelling hypothesis concerning performance functions. The hypotheses envisage: (1) to use deterministic value; (2) to pursue a stochastic approach but with an higher level of aggregation: container undifferentiated and no distinction inside loading/unloading activities, e.g. loading from dock and from shuttle are not distinct; (3) to pursue a stochastic approach with an higher level of disaggregation. The effects have been evaluated estimating global and local indicators, simulating loading/unloading operations for a set of observed vessels and simulating the trip time for a set of observed containers. Results show remarkable differences among the different hypotheses, but point out how the effects may be not negligible in short-medium term simulation scenarios (real-time management, tactical planning), whereas may be not relevant in strategical planning.

Simulation of a container terminal through a discrete event approach: literature review and guidelines for application

DE LUCA, STEFANO
2009-01-01

Abstract

Design and project appraisal of container terminals may be carried out through two main approaches. optimization or simulation. Simulation can help to achieve different aims: overcome mathematical limitations of optimization approaches, support and let computer generated strategies/policies more understandable, support decision makers in daily decision processes through a “what if” approach. Although in literature numerous efforts may be found in the field of simulation of a container terminal, most of the existing papers do not pay great attention on the model set-up, its calibration and its validation, but they are only focused on the application and/or on the comparison of design scenarios. In a container terminal may be distinguished macro-operations, operations and activities. Macro-operations are set up by operations; operations are set up by handling activities. In such a classification, dependent on the level of aggregation chosen, a performance function can be associated to an entire operation (e.g. vessel loading) or to a single handling mean (e.g. a quay crane that load a vessel). In all the cited cases, performance functions can be deterministic or stochastic and are highly influenced by containers typology (e.g. 20 feet vs 40 feet) and by their state (full or empty). Although the estimation of performance functions should be one of the main issues of all container terminal applications, such a problem does not seem to be treated deeply in most of the application existing in literature. If on the one hand, many contributions do not present any information on performance functions used, the remaining contributions carry out very simple approaches (deterministic) and/or give scanty information: on the estimation approach pursued, on experimental data used, on parameters estimated and on parameters value. In this paper, drawing on the model architecture proposed in a previous contribution by the same authors, a discrete event simulation model is proposed for the Salerno container terminal (one of the major private container terminal in the South of Italy) in order to address some of the topics introduced before and, in particular, to cope with the following issues: 1. calibration of stochastic distribution function to estimate terminal handling activities, 2. comparison of different stochastic estimation approaches ( Moment, Maximum Likelihood and Kolmogorov-Smirnov), 3. analysis of the effects of different estimation approaches (deterministic, stochastic aggregated, stochastic disaggregated) on the estimation of the whole terminal performances. The main container terminal operations have been subdivided into elementary activities and each activity has been analyzed individually. In particular, the handling means considered have been: quay cranes, yard crane and reach stackers. The activities taken into account have been:  as regard quay cranes: loading time from dock to vessel, loading time from shuttle to vessel, unloading time from vessel to dock, unloading time from vessel to shuttle;  as regard yard cranes: unloading time (to shuttle/truck), loading time (from shuttle/truck), unloading time (to stack), loading time (from stack), trolley speed (with container), free trolley speed, crane speed;  as regard reach stackers: unloading time from shuttle/truck, loading time to shuttle/truck, stacking time (to tier). All of them have been analyzed for the following container typologies: undifferentiated, 20’, 40’, 2 x 20’, full and empty. Starting from the empirical frequency distributions, different distribution functions have been hypothesised (Gamma, Normal, Weibull, LogNormal) and the corresponding parameters have been calibrated starting from real observations obtained from SCT (more than 1.000 vessels monitored) and from an integrative survey (more than 3.000 containers monitored). The calibration stage has been carried out through the implementation and the comparison of Moment, Maximum Likelihood and Kolmogorov-Smirnov estimation methodologies. For each handling means and for each container typology, deterministic values and distribution function have been obtained, have been compared and have been implemented in a discrete event simulation model. The model simulates operations of SCT, and it has been used to simulate different modelling hypothesis concerning performance functions. The hypotheses envisage: (1) to use deterministic value; (2) to pursue a stochastic approach but with an higher level of aggregation: container undifferentiated and no distinction inside loading/unloading activities, e.g. loading from dock and from shuttle are not distinct; (3) to pursue a stochastic approach with an higher level of disaggregation. The effects have been evaluated estimating global and local indicators, simulating loading/unloading operations for a set of observed vessels and simulating the trip time for a set of observed containers. Results show remarkable differences among the different hypotheses, but point out how the effects may be not negligible in short-medium term simulation scenarios (real-time management, tactical planning), whereas may be not relevant in strategical planning.
2009
9781905701049
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11386/2601100
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact