The Passenger Terminal has no capacity limit for passengers in transit and allows the turnaround of approximately passengers. Since April , a new cruise quay of metres long and metres of length for berth is in operation. Since then, the biggest and the most glamorous cruise ships from the worldwide cruise fleet are welcome here, boosting the cruise tourism in the north of Portugal. The Porto Cruise Terminal is in full operation since March and the main building has a capacity for 2. It has several amenities for transit cruise liners or turnaround cruise vessels, no limit capacity for passengers in transit and a capacity for 2 passengers in turnaround.

Since the opening of the new pier, the number of cruise ship calls and passengers at this port has been significantly increasing. ESPO: A good access to the hinterland is important to get goods to the desired destinations. Could you briefly tell us how the port is connected with the hinterland? Are you planning new projects to facilitate these connections?

The Single Gateway is a completely computerised platform, fitted with the most up-to-date infrastructures, aimed at simplifying all the procedures related to entrance or exit of heavy vehicles and their loads. This computerised platform has shown itself to be a benchmark and it has become the object of numerous study visits.

Its operation model, without waiting periods for drivers, offers speed in customer reception and assures that vehicles will stay inside the port during a minimum amount of time. The Single Gateway comprises the road haulage park and commodity control area, which gathers Customs, Concessionaries and revision and inspection equipment.

It is directly connected to the Business Management System and to the geographical information system, swiftly registering all the required information for departure or entrance. The road haulage traffic may therefore run smoothly without having to go through the surrounding cities, diminishing not only the traffic but also the related environmental pollution. How has this been of benefit for the port? It therefore works towards meeting the need of the horizontal TEN-T policy priority to provide an alternative to long distance road haulage by linking Portugal, Western France and the western coast of the UK by lo-lo and ro-ro services.

The financing opportunities are more diversified, namely within EU programmes managed directly by the European Commission. What are the other main environmental policies? At an environmental level, the port is minimising the impact of port operations activities in the physical environment in which it operates, through its permanent environmental monitoring.

Based on the literature review, the following questions were set for this simulation study. The first set of three questions relates to the quantification of the impacts of berth allocation policies and queue priorities on the time a ship spends in queue. These questions are:. In each policy above, eight different queue disciplines or priorities were considered Hansen, ; Silberholz et al. These disciplines decide which ship in queue is going to be served next, at the moment a berth becomes available.

They are given as follows:. The second set of questions is related to the formalization of elements to support the strategic positioning of port authorities. These new questions are an unfolding of the results obtained for the three previous questions, in light of different demurrage penalties for different ship types.

These secondary questions are:. In order to address the previously mentioned questions, a series of controlled experiments with two factors berth allocation policy and queue priority and several levels was developed and conducted, so as to assess their impact on the time a ship spends in queue and its major statistics mean and variance.

For this purpose, the SBL operation of a Brazilian private container terminal was modeled. The layout of this container terminal is depicted in Figure 2. The possible alternatives for the SBL operation are illustrated in Figure 3 by means of different flowcharts. They basically indicate the same sequence of activities and decisions, differing only by the factors tested.

In general terms, the ship service begins when the ship arrives at the port. Depending on congestion and priority assigned, the arriving ship may have to wait until a berth is available. After berthing, containers are unloaded loaded from on the ship. Finally, when the service is completed, the ship leaves the port. For each one of the combinations of berth allocation policies and queue priorities, 30 replications of days of operations were conducted. Statistics on the waiting time in queue, for the system as a whole and for each ship that periodically berths at the port, were collected.

It is worthwhile mentioning, as a methodological note, that the SBL model focused on ships that visit the port at least two times per year. In , from 23 ships, 14 met this criterion. They were the objects of analysis of this case study. The dependent and independent variables of the model are presented in Table 1 , which also shows their scales, dimensions, and operational definitions. The control variables, that is, the ones that had no variation for all combinations, were based on data collected in this case study. They are the mean interarrival times and the mean processing times at berth for each ship.

It is also assumed that the port resources equipments, capacities, and so on not considered in the scope of the SBL operation modeled in ARENA did not influence the statistics collected for the time spent in queue by each ship. The interarrival times for each ship were assumed to be exponentially distributed and their means are given in Table 2.

The processing times at berth for each ship were assumed to be normally distributed and their means are also given in Table 2. Four sublevels 0, 0. Several studies indicate that empirical distributions of lifts per ship adhere to normal distribution Dragovic et al. Since the container terminal is relatively small, with only two berths, another simplification in the SBL model was taken.

That is, the choice of anchoring one ship into one specific berth, rather than another, was supposed not to impact on the distance between the ship and the location point of its containers in the yard. Finally, it was also considered that the remainder port resources, which were not modeled in ARENA, did not impact on the time spent in queue by each ship.

According to Hair et al. The selected covariate was the coefficient of variation of processing times. The maximum acceptable significance level was 0. It is important to observe that this study differentiates itself from previous efforts reported in literature not only because it experimentally assesses the signs of the relationships between berth allocation policies and queue priorities, in terms of the means and variances of waiting times in queue, but also because it quantifies their main effects by means of multivariate statistical analyses, controlling them by the coefficient of variation of processing time at berth.

The case study on a two-berth container terminal and underlying assumptions serve the purpose of exploring in depth, by means of a simplified simulation model, the influence of different berth allocation policies and queue priorities in the SBL operation.

## New. Critical Factors For Berth Productivity In Container Terminal

MANOVA was performed with two dependent variables related to the mean and the variance of waiting times in queue for each one of the 14 ships and the system as a whole. In total, 3, observations were considered: 30 replications per factor combination x 4 berth allocation policies x 8 queue priorities x 4 levels of coefficient of variation of processing times.

As mentioned before, each replication consisted of a run length of 1, days of operation, from which the first days were discarded warm-up period. No outlier observations were detected among cells at the 0. Satisfactory results were also obtained for normality, homogeneity of variance-covariance matrices, linearity, and multicollinearity assumption tests. According to Wilks' lambda values Hair et al.

These results indicate strong relationships between dependent and independent variables, with a partial association power n 2 equal to 1. The estimated parameters for the impacts of berth allocation policy and queue priority, controlled for the coefficient of variation of processing times, are presented in Tables 3 and 4 addressing research questions 1 and 2.

The coefficients represent the marginal impact of other berth allocation policies "Dedicated berths according to ship size", "Single queue distributes ships to first available berth", and "Ships are allocated to the berth with the smallest expected queue size" and queue priorities "Longest processing time", "Shortest processing time", "Largest number of scheduled ship arrivals per year", "Smallest number of scheduled ship arrivals per year", "FIFO", "LIFO", and "Largest ship size" altogether with the coefficient of variation of processing time.

The intercept represents the marginal estimated means for the reference levels. Table 3 also indicates the values of F and adjusted R 2 statistics for the models that are associated to each dependent variable. The signs presented in Table 3 for the expected waiting time in queue whole system corroborate the results found by Dragovic et al. The queue priorities "Smallest ship size" and "Shortest processing time" present, respectively, null or negligible impact on the average waiting time in queue when compared to other disciplines. The same occurs under "Largest number of ship arrivals" per year.

Considering this priority, one can easily realize cf. Table 2 that the ship size is positively correlated to its processing time 0. On its turn, the ship processing time is strongly and negatively correlated to the number of arrivals per year These three queue priorities are, therefore, related to different aspects of the same operational characteristic: smaller size, shorter processing time, and higher number of arrivals per year. Considering each decision, the relative impact on each dependent variable is indicated by the modulus of the coefficients cf.

Tables 3 and 4. However, the draw of further conclusions depends not only on the aggregation level of the analysis whole system or a given ship , but also on the specific combination among the berth allocation policy, the queue priority, and the coefficient of variation of processing time.

For example, analyzing the whole system, if berths are dedicated according to ship size, the berth allocation policy will present the highest positive impact on the average waiting time in queue 4. Results should be analyzed in a similar way for all other independent variables. As exemplified, this analysis provides an overview of the structure of the average waiting time in queue for the system as a whole. However, specifics of small and large ships related to their sensitivity to variations on processing times, demurrage cost ratios, and number of arrivals per year are not yet captured.

Analyzing Table 3 , it is possible to affirm that small ships would benefit more from "Dedicated berths according to ship size" or "Single queue distributes ships to the first available berth" policies. Small ships are also less impacted by the coefficient of variation of processing times. With respect to queue priorities, it seems to be a balance between "Shortest processing time" and "Smallest ship size" in terms of their impact on the average waiting time.

On the other hand, when large ships are considered, one can realize that they would benefit more from "Single queue distributes ships to the first available berth" or "Ships are allocated to the berth with the shortest expected queue time" policies. As regards the assignment of a given queue priority, however, the picture is not so clear.

It is also possible to affirm that large ships are more impacted than small ones by the coefficient of variation of processing times. Figure 5 presents the marginal expected values of the average waiting time in queue for each ship at each one of the combinations between berth allocation policies and queue priorities.

- Bibliographic Information.
- Port Productivity | Global port productivity fell in , as call size continued to grow?
- Future trends on container handling systems.

These expected values were evaluated at the mean level of the coefficient of variation of processing time. It is clear that these countervailing forces within berth allocation policies and queue priorities for each ship should be taken together into account. More precisely, they should be weighted not only by the number of arrivals per year, but also by the demurrage cost of each ship, in order to determine what combination would lead, in fact, to the smallest total demurrage cost for the whole system.

However, an analysis simply based on the marginal expected values of the average waiting time in queue for each ship may not reveal the full picture about total demurrage costs. Therefore, a probabilistic analysis is deemed necessary. Figure 6 presents the marginal expected values for the variance of the waiting time in queue for each ship at each one of the combinations of berth allocation policies and queue priorities.

## Critical Factors for Berth Productivity in Container Terminal

These values were also evaluated at the mean level of the coefficient of variation of processing time and calculated based on the coefficients in Table 3. More precisely:. The best cobi nation is the one that leads to the smallest T DC. Before proceeding with the discussion on how this expected demurrage time can be determined based on Table 3 , some additional considerations on the choice of the gamma assumption for waiting times in queue must be made first.

The choice of the gamma distribution is due to its main properties. Also, the gamma distribution is related to several other distributions, thus making this analysis more robust. With respect to the determination of the conditional expected demurrage for the gamma waiting time in queue, Tyworth et al. They demonstrated the relative simplicity of these solutions and discussed some practical considerations as regards to their implementation in electronic spreadsheets, like MS-Excel.

In their analytical form and as a spreadsheet function, these solutions are, respectively, given by:.

The demurrage cost ratio indicates how many times the demurrage cost per hour of a large ship is greater than the demurrage cost per hour of a small ship. The TDC was calculated for each one of the combinations of berth allocation policies and queue priorities. The coefficient of variation of processing time used was 0. In this case, a clear priority would be given to small ships, so as to favor the cost performance of the port system as a whole. Ships numbers 5 and 8 would be exceptions, since they are large ships with short processing times.

Priority, however, should be assigned to the ships with the largest number of visits per year. Under this criterion, large ships with large number of visits per year would be favored, thus contributing to reduce the total demurrage cost of the system. Ships numbers 6, 7, and 8 meet this criterion. Results presented in Figure 7 address the research questions 4 and 5, and indicate the contradictions with respect to the aggregate results in Table 3.

One may easily note that this combination is a mixture of the three other combinations depicted in Figure 7 , since the specifics of each ship operation are not being considered in the SBL problem. The SBL problem is complex, because of the different sizes of ships, different interarrival times, and different processing times at berths. By means of a simulation of a two-berth container terminal, developed in ARENA, this case study assesses the impact of different berth allocation policies and queue priorities on the waiting time spent in queue and on total demurrage costs.

This paper deals with theory and practical decision-making. In comparison with previous literature, two elements, detailed next, constitute its contribution. The first element is the simultaneous consideration of four different possible berth allocation policies in parallel to eight possible queue priorities. By means of simulation, this study provides experimental confirmation of the available evidence in literature on the queue priority assignment to ships with the shortest processing time, so as to improve the overall performance of the port, in terms of expected waiting time in queue.

The second element is the incorporation of the demurrage cost for each type of ship to determine the most adequate combination of berth allocation policies and queue priorities. The outputs of the simulation, generated for each ship, are further analyzed and weighted in terms of demurrage probabilities and costs in order to deal with the apparent contradictions between analyzing the SBL operation as a whole and the SBL operation as a weighted sum of the specifics of each ship.

A limitation of this study is related mainly to the underlying assumptions adopted at the design and execution of the experiments: the choice of berthing one ship rather than another into one specific berth is considered not to impact the distance to the location point of its containers in the yard; the impact of the number of lifts and quay cranes is considered to be embedded within the average processing times for each ship; and it is also considered that the scope of the SBL operation modelled in ARENA does not impact the time spent in queue by each ship.

### Recommended for you

Another limitation is related to the small size of the terminal studied, thus making it more difficult to generalize the conclusions generated to larger operations. Nevertheless, despite such limitations, the simplified model and assumptions in this case study have the advantage of allowing an in-depth understanding of the primary effects of berth allocation and queue priority on port performance and their impact on total demurrage costs.

Future research should be conducted under less restrictive assumptions. Port management and operations. London, Informa. Allocation of ships in a port simulation. A queuing network model for the management of berth crane operations. Simulation and the lean port environment. Analysis of operations at the Kaohsiung port new mega container terminal. The Business Review, 4: Maintaining Singapore as a major shipping and air transport hub.

In: Toh T Ed. Singapore University Press, Singapore, pp.