论文部分内容阅读
为减少由网络分解造成的网络分析精度损失,基于系统科学中节点删除的方法,提出一种度量节点间连通性影响的指标——相对连通系数.以该指标为依据,应用主成分分析方法提取与目标相关的节点来生成子网的方式建立了新的网络分解方法,在解空间损失较小的情况下降低了网络分析的计算复杂性.求解最短路径的试验表明:该方法可有效控制网络分解造成的精度损失;网络规模压缩至原有的20.12%,而最大误差为13.85%;计算时间由秒级降至百毫秒级.
In order to reduce the loss of network analysis accuracy caused by network decomposition, based on the method of node deletion in system science, a metric to measure the influence of inter-node connectivity is proposed, which is based on this index and extracted by principal component analysis And the target-related nodes to generate sub-networks, a new method of network decomposition is established, which reduces the computational complexity of network analysis under the condition of small solution space loss.The experiment of solving the shortest path shows that this method can effectively control the network The accuracy of the loss caused by the decomposition; network size compressed to the original 20.12%, while the maximum error of 13.85%; calculation time from second to 100 milliseconds.