基于改进遗传算法的泊位岸桥协调调度优化

Integrated berth and quay-crane scheduling based on improved genetic algorithm

针对集装箱码头资源调度不合理造成资源浪费的问题,在考虑岸桥装卸成本的基础上,以在港集装箱船总的作业成本最小为优化目标,建立了基于非线性混合整数规划的泊位岸桥协调调度优化模型。为使模型更加接近码头操作的实际情况,模型假设船舶装卸时间依赖于为其分配的岸桥数。采用基于可拓关联函数的改进遗传算法对模型进行求解。改进算法强调了不可行解的重要性,用可拓关联度来衡量种群中不可行解的优劣程度,通过在种群迭代中始终保持一定数量的不可行解来维持种群多样性,从而克服传统算法局部搜索能力较差的缺陷。数值实验验证了模型和算法的可行性和有效性,与不考虑岸桥装卸成本的模型相比,能够有效减少港口资源的浪费。

A strategy for integrated berth and quay-crane scheduling was proposed to cope with unreasonable allocation of port resources in container terminals. First, a nonlinear mixed integer programming model which aims at minimizing the port operational cost was presented. And the loading and unloading cost of quay-crane was considered in the objective of our model. To make the model more realistic, the handling time of a vessel was assumed to depend on the number of assigned quay-cranes. Second, an improved genetic algorithm based on extenics dependent function was used to solve this model. In this algorithm, infeasible solutions play an important role. They were evaluated by their extenics dependent degrees. Some infeasible solutions were always contained in the population to maintain the diversity of the population. This improved local search ability of traditional genetic algorithm. At last, the effectiveness and efficiency of the proposed model and algorithm were testified by several test instances. Compared with the model without considering the loading and unloading cost of quaycrane, the waste of resource is effectively reduced.