摘要
将摄动对偶单纯形算法和亏基算法相结合,不仅为亏基的对偶单纯形算法提供了一个初始对偶可行基,还能有效的克服退化现象,并用matlab编制程序对22个标准NETLIB问题进行数值试验数值,结果表明,该算法可在总体上减少迭代次数,极具竞争性.
By combining the perturbation dual simplex algorithm with deficient-basis algorithm, an initial dual feasible basis can be provided for dual simplex algorithm of deficient basis. Computational difficulties yielding from degeneracy can be overcome. Numerical test values can be done on the twenty-two standard NETLIB problems with the matlab programming. Preliminary computational experiments show that it requires less iterations overall, and is hence very competitive.
关键词
线性规划
亏基
摄动
退化
对偶单纯形算法
linear programming
deficient basis
perturbation
degeneracy
dual simplex algorithm