Research Of Dynamic Berth Allocation Planning In Container Terminal

With the development of economy, container terminal throughput has increaseddramatically and a number of container terminals have exceeded the capacity which isdesigned originally. However, many container terminals can’t upgrade the hardware facilities,expand the yard area and increase shoreline length due to the rising land prices and expensiveterminal equipment. In order to increase the port capacity, the operational efficiency ofcontainer terminal must be increased. Therefore, it has become the focus of port study whichis extremely difficult. The shipping companies must meet the demand of their users. They areboth determined by berth allocation directly. So berth allocation problem (BAP) attracts theattention of more and more researchers and companies.Through studying the allocation process of container berths, the paper takes the shortesttotal time of the ships in port as objective function and set up various constraints of the modebased on actual situation of the berth allocation. Finally the paper establishes a berthdynamically allocated (DBAP) model to achieve dynamic allocation. It has importantpractical significance.In order to solve the model, the improved multi-coded genetic algorithm, which isimproved and innovated on fitness calculation encoding and select operation, is designed. Thearticle also uses Matlab design to develop solving program of berths dynamic allocationmodel based on multi-layer coded genetic algorithm. Then a numerical experiments isconducted and a group of reasonable algorithm operating parameters are obtained. When anumerical example is used on a terminal actual production data, the optimal solution obtainedby the design algorithm reduces on average by10%or even15%in better results, the totaltime in port, than the traditional berth allocation strategy that first comes and first serves.The article also designs UML modeling method based on DEVS for simulation modelingof complex systems. This method combines the features of the DEVS, which is hierarchicaland modular, and the UML, which is dynamic description and object-oriented. This method isused to analyze and design berth simulation system achieved by eM-Plant. When comparingoptimal solution obtained by genetic algorithm, from simulation and optimization system onthe HP berths operating system, with the solution generated by first come first service strategy,it verifies the feasibility of the algorithm and draws the conclusion that the parameters of theberth of berth operating system are optimized through using the program obtained by theoptimization algorithm.The article also designs a dynamic berth allocation scheme based on the rolling window technique achieved in simulation model, so that the simulation model can continuouslysimulate the real situation and meet daily operations of the terminal and normal schedulingproduction in emergency situations. This study can provide the decision basis in berthsallocation, thereby improving the efficiency of terminal operations.