一种新的Lagrange乘子法

A New Lagrange Multiplier Method

  • 摘要: 基于Lagrange乘子法中将与不等式约束相关的乘子定义为原乘子的正定函数,用同样的方法处理不等式约束和等式约束的构想,构造了一种新的Lagrange乘子法.分析了该算法的收敛性,并利用LaSalle不变集原理揭示了算法稳定机制及如何减弱收敛条件和扩大收敛域.分析表明,算法在稳定因素和不稳定因素的综合作用下获得最优解.

     

    Abstract: Inequality constraints were handled with the similar method as their equality counterparts in Lagrange multiplier method,while the multipliers associated with inequality constraints were written as a positive definite function of the originally-defined multipliers.Based on this,a new Lagrange multiplier method is proposed and its convergence rigorously analyzed.The underlying mechanism that leads to the algorithmic convergence is uncovered,and some measures for relaxing the convergence conditions and enlarging the attractive domain are discussed by the LaSalle invariance principle.The result showed that the algorithm attained optimal solution both by stable and unstable factors.

     

/

返回文章
返回
Baidu
map