基于2类决策和改进蚁群算法的多CCD调度  

Research on multi-CCD sensor planning based on two classes of decisions and improved MMAS

作　　者：赵宇[1] 王先培[1] 乐意[1] 张其林[1] 杜双育[1] 

机构地区：[1]武汉大学电子信息学院,武汉430079

出　　处：《仪器仪表学报》2012年第2期474-480,共7页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(50677047);湖北省自然基金(2010CDZ051);湖北省教育厅基金(D20092505)资助项目

摘　　要：多CCD系统的调度要求同时兼顾调度质量与实时性。以无人值守变电站视频监视系统为研究对象,在考虑系统特性的基础上,以目标与CCD间距、节点载荷、优先级和模糊预测因子为指标,使用模糊决策动态调整预测区域半径,并建立基于2类决策的快速调度模型。对多目标调度的求解,研究了改进的最大最小蚁群算法,通过引入合适的收敛性判据和状态转移规则,避免了过多使用启发式算法造成的决策耗时,提高了决策的效率和算法的全局搜索能力。最后进行了系统仿真和真实实验,实验结果表明该方法可减少激活节点的数目,特别是对于较复杂的多CCD-多目标调度,能同时保证算法具有较好的实时性和求解性能,验证了模型的有效性。The objects of muhi-CCD ＆ multi-task scheduling are quality and real-time property. Considering the characteristics of the remote-image-monitoring system for unmanned substation, a rapid-scheduling model based on two classes of adaptive decisions is founded, which includes four parameters： the distance between CCD and target, the load of nodes, the mission priorities and the fuzzy-forecast factor. The predicted region radius is dynamically ad- justed through fuzzy decision. An improved MMAS algorithm is then introduced for multi-task scheduling solution, which adopts proper termination criterion and transfer rules to improve the efficiency and global search capability of the algorithm. Experiment data from simulations ＆ actual tests show that the proposed method can reduce the number of activated CCDs and provide more satisfactory performance for CCD-planning, especially for multi-CCD ＆ multi- task scheduling.

关 键 词：多CCD调度 模糊预测因子 快速自适应决策 最大最小蚁群算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]