基于改进粒子群算法的航班降落调度问题研究

航班降落调度是多约束的整数规划问题,也是组合优化中的典型NP难题之一。针对传统粒子群算法易陷入局部最优和求解组合优化问题能力不强的缺点,提出改进策略。改进的算法通过禁忌搜索策略生成初始粒子群以满足航班调度多约束的限制,引入遗传算法的交叉变异等操作以增强粒子群间信息交流能力,重新定义惯性权重和加速度因子,以增强粒子的多样性和收敛性,加入进化交叉算子以增强粒子的局部寻优能力。通过与修补遗传算法、罚函数遗传算法进行对比,经多次运行后的结果表明,提出的改进算法的局部搜索能力和全局搜索能力都得到增强,同时也验证了算法的可行性和有效性。

Evolutionary Computation for Realizing Distillation Separation Sequence Optimization Synthesis

Evolutionary algorithm is applied for distillation separation sequence optimization synthesis problems with combination explosion. The binary tree data structure is used to describe the distillation separation sequence, and it is directly applied as the coding method. Genetic operators, which ensure to prohibit illegal filial generations completely, are designed by using the method of graph theory. The crossover operator based on a single parent or two parents is designed successfully. The example shows that the average ratio of search space from evolutionary algorithm with two-parent genetic operation is lower, whereas the rate of successful minimizations from evolutionary algorithm with single parent genetic operation is higher.