Optimising Container Transfers at Multimodal Terminals
The seaport terminals have changed dramatically after the introduction of containerization. These changes include alterations to the storage area！ the
introduction of specialised container handling equipment and storage methods (stacking abilities). The role of a multimodal container terminal (MCT) is to ensure a smooth transfer of freight between the two modes. Such freight may be in containers！
flat trays！piggyback (trailer on flat wagon)！or roadrailers (trailers capable of
road and rail movement without requiring rail wagons). The main factors influencing terminal operating performance are as follows. Operating strategies; physical layout！ ship and train plans; management/ work practice; ship and train reliability; pick-up-delivery cycle times; lifting equipment and customer requirements. Some of these factors will now be discussed below.
Two main operating philosophies for the loading and unloading of containers are the random access system and the use of skeletal trailers. Under the random access system！ customers deliver/pickup containers directly to/from a train or to/from ground storage. This is the method commonly used in Australian and European intermodal freight terminals. The skeletal trailer system is mainly used in North America and is based on the use of a dedicated fleet of skeletal trailers which are used to pick-up containers directly from trains. [2J Those trailers are then moved to a trailer storage area ready to be picked up by individual customers. The reverse process is followed when loading on to the train. This study uses the random access system operating philosophy.
The equipment available to handle containers in the intermodal terminal are of three main types: gantry cranes (rail mounted or rubber tyre); side loaders (forklifts and reachstackers) and straddle carriers (rubber tyre). The choice of equipment will depend on container throughput！ operating strategy！ physical
operating space！ track layout and degree of standardisation in container sizes and types. Each type of equipment has different capital cost！ land requirements for
operating purposes and pavement strength requirements.
Overall transit times！ reliability of delivery times and costs are the main factors influencing mode choice in the freight transport sector. Users of intermodal terminals have as their main requirements: reliability of delivery times！ container
pick-up and delivery cycles which are delay free！ and the ability to monitor the
progress of their consignments (i. e.！ real time information regarding container
location and estimated arrival times).
Whichever technology is applied！ it has to be taken into consideration that
the“container transport” system consists of a number of sub-systems！ the capacities
of which need to be well harmonised in order to prevent bottle-necks within the
transport chain. It is absolutely essential to meet this demand！ not only with regard
to investment in facilities！ but also with regard to the operational management.
The ideal situation is for containers to be transferred to the berth before the arrival of the ships to reduce the port time！ and then to be stowed on the ships
by the shore cranes. If the import containers are not on the top of the ship then the other containers should be unloaded and restowed. Thus！ many containers on the
dock will causes delays.
Daganzo has analysed the effect of crane operations on ship service at port terminals. Daganzo has also calculated the maximum berth throughput during periods of congestion. He found that the average ship delay can vary considerably with the crane operating strategy Peterkofsky and Daganzo used a crane allocation scheme to minimise the cost that ships incur in port. Taleb-Ibrahimi has analysed the effect of handling and storage strategies for seaport terminals. Noritake and Kimura have analysed the movement of ships in a port and have specified the costs spent at public wharves. In their later study Noritake and Kimura proposed a method to determine the optimum allocation and size of ports in a country from a national economical point of view.
When a container vessel calls to port！ the containers on board must be unloaded
and stored at the port until they are transported further by rail or road. The containers must be stored in a manner so as to minimize the amount of handling needed to place a container in the storage area and to remove it when needed. Therefore, the problem being investigated is the minimization of the total throughput time which is the handling time for all the containers from ships at berth and the transferring time of the containers to the MCT. When dealing with export containers！ the problem
would be reversed; that is！ the minimization of the handling time of the containers from first arrival at the port until the ship carrying the containers departs from the port. When a vessel arrives and its cargo is .unloaded！ the stevedoring company
will receive information about where some of the containers are to be transported.
The containers that are remaining must be placed in storage areas until they are needed. The company does not know when or in what order the containers will be called for loading or unloading. Therefore！ they must stack the containers in a
manner！ so as to minimize the time taken to retrieve a container by considering the storage area constraints. In the case of exports！ the stevedoring company
usually knows when a container will depart as it arrives. The stevedoring company charges a fee for containers that are delivered too early in respect to the departure time and after cut-off times no containers are received.
The Brisbane Multimodal Terminal (BMT) works by removing import containers from the marine container terminals by trucks and then transferring them on to container wagons at the BMT. Export containers arriving by rail are transferred to the marine container terminals by BMT trucks. Empty container wagons are prepared for the next trip and stored at container parks which are adjacent to the Brisbane Multimodal Terminal. Reachstackers and forklifts can handle 20 feet or 40 feet containers at the terminal to load and unload the rail wagons and transport containers between the wagons and the BMT trucks.
The Brisbane Multimodal Terminal eliminates costly shunting！ and thus！ saves
time and money for importers and exporters in the base of operation. There are currently two container terminals at the Port of Brisbane！ with a total length of
1！ 300m. The terminals are owned by the Port of Brisbane Corporation but are leased to two stevedore companies. One container terminal (Berths 2 and 3) is leased and operated by Australian Stevedores and the other container terminal (Berths 4 and 5) is leased and operated by Conaust Ltd.
Shore cranes are used to lift the containers on and off container vessels. The containers are unloaded into the marshalling area where they wait until forklifts or highstackers move them to the storage areas or to awaiting trucks for transportation to the Brisbane Multimodal Terminal. Each container terminal has
storage areas but remote storage areas are also located on Fisherman Islands and at other locations around Brisbane！which include Murrarie！Acacia Ridge！
The trial data set used for the solution and subsequent sensitivity analysis is detailed below. It takes an average of 2. 7 minutes for each crane to unload one container. From the marshalling area！an average of 15% of containers are transferred to the BMT trucks！ an average of 5 % of containers are transferred to the empty container storage area and the remaining 80 % of containers are transferred to stored in the two storage areas. It takes ten minutes per container to move from the marshalling area to the empty container storage area.
When using Berth 1！ the containers to be stored are moved to Storage Areas ！2 and empty storage area. Each of the storage areas is divided into four 1
！ and each section has a different traveling time but all sections have sections
a capacity of 500 containers/stacks. The maximum level of stack is three. The distribution of handling time of highstackers is Erlang. The distribution of
l. The traveling times are normally dis-handling time of shore crane is norma
tributed. Mean traveling times from the marshalling area to Section 1 takes 3 minutes！to Section 2 takes 3.5 minutes！to Section 3 takes 4 minutes and
r. To move a container from the to Section 4 takes 6 minutes per containe
r. marshalling area to the BMT trucks takes an average of 3 minutes per containeThe containers moved to the BMT tracks are transported to the BMT and each container takes an average of 6 minutes. Eventually！ all the containers in
l. From the storage will be moved to the BMT to be distributed by road or rai
！ an average of 73.2， will total of containers that were unloaded from ships
be distributed by rail and the other 26.87， by road.
At the present container terminal！ the number of shore cranes available for
use is two！ the number of highstackers (include forklifts) is ten and the number of trucks is six. The problem was solved by using GAMS for different time periods. Sensitivity analysis was performed with the same information but changing the number of shore cranes highstackers！ and trucks available for use.
This study has been confined to the basic elements of the overall investment planning problem related to the expansion of the system. Improvements in operational methods are beyond the scope of this study. This model could be further investigated
by carrying out studies into the effects brought about by improvement in operational methods.
Investments in multimodal terminals are very costly and the technical progress of the equipment used gives them a much shorter life than they had in the past. In order to obtain maximum benefits it is usually necessary to combine a number of investment strategies into a coherent and complementary package of capital expenditure projects. For example！ the investment in
terminal infrastructure to allow faster loading/unloading of ships and trains.
The problem being investigated is the minimisation of handling and traveling time of containers from the time the ship arrives at port until all the containers from that ship leave the port. If dealing with export containers the problem would be reversed. That is！ the handling and traveling time of the containers from when they first arrive at the port until the ship carrying the containers departs from the port. This mathematical model can be used as a decision tool in the context of investment appraisals of multimodal container terminals. Long-time data collection should be carried out before the implementation of the model. In the optimisation of the port system through this type of mathematical model！ several parameters are
involved in the phenomena which influence the optimisation results. A more detailed study may be undertaken to analyse the effect of these parameters on the improvement of port capacity in the long-term. The model assumes that equipment is available every time it is needed.
A cost-benefit analysis must be performed before any implementation is considered. To make the cost-benefit analyses results more flexible all analysis should be carried out to determine the sensitivity of the optimal solution to changes in variables. This would provide port planners with a mechanism for the continual updating of the optimal solutions based on any new estimates of these parameters.
In addition！ a comprehensive hinterland analysis within the national context will provide more comprehensive data for estimating the future demand on any seaport system. Future studies are needed on the alternative means of increasing seaport efficiency by improving utilization of the present capacity. Such a study might cover better port planning methods！ investments for increasing the capacity of the lagging segments of the seaport system！ and means of better utilization of present
1.2 New Words and Expressions
1. cargo n.(车、船、飞机等运输的)货物