模糊系统由一般输入变量、推理规则和实际输出组成,推理规则总数会随着输入变量的增多而呈指数形式急剧上升,容易引起规则爆炸.为此,降低规则总数是必要的.本研究首先对二叉树型分层广义混合模糊系统进行结构分析;其次,依据混合推理规则及计数公式针对分层前与分层后的规则总数进行比较和分析.结果表明,分层后系统内部规则总数的增长速度远远小于不分层的规则总数的增长速度,从而可避免规则爆炸现象.

针对超大型火箭集装箱协同吊装的实际需求,研究和实现了一种基于虚拟现实技术的协同吊装模拟训练系统,首先阐述了系统的总体架构;然后基于开源物理引擎ODE (Open Dynamic Engine),提出了一种基于多刚体动力学模型的钢丝绳模拟方法;在深入分析协同吊装作业过程和特点的基础上,研究和建立了一种协同吊装动力学模型;针对协同吊装过程中各仿真计算机的视景同步显示问题,提出了一种有效的视景同步显示方法.实验结果表明,该系统可以逼真的模拟船载起重机对超大型火箭集装箱进行协同吊装作业的全过程,能够满足日常吊装训练需求,节省训练成本,提升训练效果.

The increase in world trade has spurred the tremendous growth in cargo containerization through container ports, which has put pressure on the port to make facilities more efficient to cope with this dynamic challenge. Without an effective operation planning of yard cranes, trucks may have to wait in the yard, and consequently QCs (quay cranes) will be idle in waiting for trucks. Therefore, to achieve high productivity, the use of yard cranes should be well planned.In this paper, the yard crane scheduling problem seeks an optimal schedule for a given collection of container flows, each of which requires a known sequence of handling operations in storage blocks, where only one container can be mounted by each yard crane at a time. Considering yard crane problem's constraints, for example, the precedence constraints among the operations and the prevention of conflicts between container flows, we formulate a yard crane scheduling problem mathematical model. To solve this problem, we propose a Genetic Algorithm (GA) approach, which uses operation-based representation, partial schedule exchange crossover and job-pair exchange mutation. The proposed Genetic Algorithm method is compared with heuristics: Shortest Processing Time (SPT) and Longest Processing Time (LPT) in the different instances of yard crane scheduling problem. Comparative analyses show the proposed GA provides optimal solutions in all the cases.