改进的免疫优化算法对动态约束多目标问题的应用

基于进化理论的动态多目标优化算法极易陷入局部最优,跟踪动态Pareto有效面的速度及效果较差。基于免疫系统机理提出一种改进的免疫优化算法(DMIOA)用于动态约束多目标问题求解。算法通过抗体浓度及其支配度设计抗体与抗原亲和力,随机约束选择算子提高算法约束处理能力,环境识别算子自适应判断环境变化,根据识别结果以不同的方式产生新环境的初始抗体群。数值实验中,将DMIOA应用于两种动态标准测试问题及飞机减速器参数动态设计问题的求解,结果表明:DMIOA能快速跟踪动态Pareto有效面,且在各环境所获面分布均匀,具有较好的实际问题求解能力。

电力系统网络重构的多目标双层优化策略

提出一种节点重要度综合评价新方法,并在此基础上发展了网络重构多目标双层优化模型。在上层优化模型中,以最大化系统可用发电容量确定发电节点恢复顺序;在下层确定恢复路径的多目标优化模型中,以最大化恢复路径的平均重要度和最小化路径的充电电容为目标,并考虑了线路恢复时间对机组恢复的影响。提出一种基于细胞生理特性的连续动态Pareto多步凋亡优化策略(MSAOS),其将整个电力系统网络重构划分为多步,每步的优化问题规模和难度都较小,且可在一定程度上避免维数灾问题;利用该策略进行多目标网络重构优化,可以在兼顾全局最优的基础上实现发电节点恢复路径优化的分段最优,并能以分支图的形式展示恢复优化过程中的网络重构策略集。利用模糊决策(FDM)方法优化出最终网络重构策略。以新英格兰10机39节点系统和广州电力系统为例,说明了所发展的模型和方法的基本特征。

Multi-objective optimization for deepwater dynamic umbilical installation analysis

We suggest a method of multi-objective optimization based on approximation model for dynamic umbilical installation. The optimization aims to find out the most cost effective size, quantity and location of buoyancy modules for umbilical installation while maintaining structural safety. The approximation model is constructed by the design of experiment (DOE) sampling and is utilized to solve the problem of time-consuming analyses. The non-linear dynamic analyses considering environmental loadings are executed on these sample points from DOE. Non-dominated Sorting Genetic Algorithm (NSGA-II) is employed to obtain the Pareto solution set through an evolutionary optimization process. Intuitionist fuzzy set theory is applied for selecting the best compromise solution from Pareto set. The optimization results indicate this optimization strategy with approximation model and multiple attribute decision-making method is valid, and provide the optimal deployment method for deepwater dynamic umbilical buoyancy modules.