摘要
将有向图中的每条边对应一个决策变量,在求解两点间的路径时,这些决策变量满足基尔霍夫约束关系。决策变量可以分为独立的和不独立的两部分,分别对应独立变量神经网络和不独立变量神经网络的状态,这些神经网络的状态代表了最短路径的解。不独立变量神经网络的状态由独立变量神经网络的状态线性组合而成,给出了独立变量神经网络方程。
When each decision variable corresponds to a edge of a directed graph,the decision variables to solve the shortest path problem must submit to a constraint that called Kirchoff's constraint.They can be classified into independent and dependent that corresponding to the states of independent variable neural network and of dependent variable neural network respectively,and these states represent the solution of the shortest path problem.The states of dependent variable neural network can be solved by linear conbining the states of the independent variable neural network,and the formulation for independent variable neural network is given.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第30期29-31,共3页
Computer Engineering and Applications
基金
国家863高技术研究发展计划基金项目(编号:2001AA14033)
关键词
最短路径
神经网络
基尔霍夫约束
独立变量
shortest path,neural network,Kirchoff's constraint,independent variables
