KKt

基本解釋

By reformulating the KKT system as a constrained equation, the algorithm generates the search direction by solving a linear equation at each iteration.
通過(guò)將問(wèn)題的KKT系統(tǒng)轉(zhuǎn)化成一個(gè)約束方程，算法在每步迭代只需解一個(gè)線性方程組即可得到搜索方向。
Under some suitable conditions, theoretical analysis shows that the sequences generated by the method converge globally to the KKT point of the point.
理論分析表明在一些假設(shè)條件下，這一方法產(chǎn)生的序列全局收斂于問(wèn)題的KKT點(diǎn)。
Methods Learning Motivation Factors Test (KKT) was conducted on 1,186 middle school students.
方法采用學(xué)習(xí)動(dòng)力因素測(cè)驗(yàn)(KKT)對(duì)1186名中學(xué)生測(cè)試。
The techniques, described here, are suitable to other systems such as KKT and SQD.
所用方法亦可用于研究KKT及SQD等結(jié)構(gòu)線性系統(tǒng)的最佳向后擾動(dòng)分析。
Then we begin to optimize other lagrange multiplications until all multiplications are accord with the KKT term.
我們?cè)賹?duì)其他Lagrange乘子做最小優(yōu)化，直到所有乘子都符合KKT條件時(shí)，目標(biāo)函數(shù)達(dá)到最小，算法結(jié)束。