Berth Allocation in Container Ports: A Particle Swarm Optimization (PSO)-Based Approach
Berth Allocation in Container Ports: A Particle Swarm Optimization (PSO)-Based Approach
复制标题
DOI：
10.1061/9780784413067.165
复制
发表时间：
2013-08
期刊：
影响因子：
--
通讯作者：
Mohammad Joharianzadeh;Abbas Babazadeh;A. Fakher
中科院分区：
文献类型：
--
作者：
Mohammad Joharianzadeh;Abbas Babazadeh;A. Fakher
以提高效率为目标的集装箱港口管理是沿海运输工程中的一个复杂问题。鉴于集装箱船舶规模稳步增长的趋势，更灵活的泊位分配规划是强制性的，特别是在繁忙的枢纽港口，各种规模的船舶停靠。集装箱码头泊位分配问题是指将泊位分配给到港船舶，使船舶的总停泊时间最小。该问题被描述为一个混合整数规划模型，假设装卸船舶的停泊地点和开始装卸时间变量为整数。由于模型的组合性质，这些假设使得模型的求解变得困难。最近，针对该问题设计了一种遗传算法(GA)亚启发式算法，并在一些测试实例上进行了测试。本文考虑了一种更灵活的BAP数学模型，其中泊位和开始时间的变量为实数。本文的目标是将粒子群优化算法(PSO)的亚启发式算法应用到该模型中。通过数值算例对算法进行了实现和测试，考察了模型的性能，并与遗传算法进行了比较。结果表明，具有实变量的船舶靠泊位置和开航时间问题的解与具有整数的问题的解有很大不同。此外，粒子群算法在较少的计算时间方面优于遗传算法。
Management of container ports, with the aim of increasing the efficiency, is considered as a complex problem in coastal transportation engineering. In view of the steadily growing trend of the container ship sizes, more flexible berth allocation planning is mandatory, especially in busy hub ports where ships of various sizes are calling. The berth allocation problem (BAP) in container terminals is defined as berth allocating to the incoming ships so that the total elapsed time of the ships is minimized. The problem is formulated as a mixed integer programming model, assuming that the variables of berthing locations and start times of handling the ships are integers. The assumptions make the model difficult to solve on account of its combinatorial nature. Recently, a genetic algorithm (GA) metaheuristic has been devised for the problem and tested on some test examples. In this paper, a more flexible version of the BAP's mathematical model is considered, in which the variables of berthing locations and start times are real numbers. The goal of this paper is applying the particle swarm optimization (PSO) metaheuristic to this model. An algorithm is implemented and tested by numerical examples, investigating the model's properties and evaluating the PSO against GA. The results showed that the solutions of the problem with real variables of berthing location and start time of ships are enough different from those of the problem with integers. Also, the PSO outperforms the GA in the sense of less computational times.