A multi-objective scatter search for the ladle scheduling problem
A multi-objective scatter search for the ladle scheduling problem
复制标题
DOI：
10.1080/00207543.2014.939238
复制
发表时间：
2014-11
影响因子：
9.2
通讯作者：
Yuanyuan Tan;T. Cheng;Min Ji
中科院分区：
文献类型：
--
作者：
Yuanyuan Tan;T. Cheng;Min Ji
我们考虑钢包调度问题，它可以被视为具有半软时间窗和调整时间的车辆路径问题。该问题涉及根据给定的炼钢调度计划分配钢包来服务钢水，并确定服务过程后空钢包的修改操作。此外，结合钢水处理时间的可控性，问题的另一个方面是考虑到实际中的工艺限制来确定服务启动时间。我们提出了一种非线性数学规划模型，其目标是最小化钢包的占用率和最大化满足软窗口的满意度。为了解决多目标模型，我们通过重新设计SS的公共组件并结合多样化生成器、组合方法和多样化标准来开发一种新的分散搜索（SS）方法，以对搜索空间进行广泛的探索。我们使用中国一家知名钢铁生产厂的三个实际实例，对所提出的方法与多目标遗传算法和实际生产中采用的手动调度的性能进行了分析和比较。计算结果证明了所提出的 SS 方法解决钢包调度问题的有效性。
We consider the ladle scheduling problem, which can be regarded as a vehicle routing problem with semi-soft time windows and adjustment times. The problem concerns allocating ladles to serve molten steel based on a given steelmaking scheduling plan, and determining the modification operations for the empty ladles after the service process. In addition, combining the controllable processing time of molten steel, the other aspect of the problem is to determine the service start times taking into consideration the technological constraints imposed in practice. We present a non-linear mathematical programming model with the conflicting objectives of minimising the occupation ratio of the ladles and maximising the degree of satisfaction with meeting the soft windows. To solve the multi-objective model, we develop a new scatter search (SS) approach by re-designing the common components of SS and incorporating a diversification generator, a combination method and a diversification criterion to conduct a wide exploration of the search space. We analyse and compare the performance of the proposed approach with a multi-objective genetic algorithm and with manual scheduling adopted in practical production using three real-life instances from a well-known iron–steel production plant in China. The computational results demonstrate the effectiveness of the proposed SS approach for solving the ladle scheduling problem.