基于最优步长的改进严格收缩PRSM算法

IMPROVED STRICT SHRINKAGE PRSM ALGORITHM BASED ON THE OPTIMAL STEP SIZE

为求解三块变量的线性约束最小化模型,基于最近提出的改进的严格收缩可分离的凸最小化模型分裂法(MSC-PRSM),改变了原算法中γ的取值方式,令新算法中的γ不再取固定值,而是采用最优步长γk进行迭代,并扩大了原算法中参数的取值范围,从而保证了算法的收敛性,证明了新算法的收敛性,并通过数值实验验证了其有效性。

In order to solve the linear constraint of three variables variable minimized model,based on the recently proposed improved strict shrinkage separation convex minimized model splitting method(MSC-PRSM),changed the value in the original algorithm,the new algorithm no longer take fixed value,but the optimal step iteration,and expand the range of the parameters in the original algorithm,thus ensure the convergence of the algorithm,proved the convergence of the new algorithm,and verified its effectiveness through numerical experiments.