对P*(κ)线性互补问题提出了一种自适应全-Newton步不可行内点算法.算法是对Mansouri等人(H.Mansouri and M.Pirhaji in Journal of Operations Research Society of China 1:523-536,2013)提出的单调线性互补问题的自适应不可行内点算...对P*(κ)线性互补问题提出了一种自适应全-Newton步不可行内点算法.算法是对Mansouri等人(H.Mansouri and M.Pirhaji in Journal of Operations Research Society of China 1:523-536,2013)提出的单调线性互补问题的自适应不可行内点算法的推广.在算法的每一次迭代中,障碍校正参数θ的取值并不固定,它总在1/(51n(1+4κ)2)和1/(14n(1+4κ)2)之间取满足算法要求的最大值,使得算法快速收敛于问题的一个ε-近似解.展开更多
Based on the generalized Dikin-type direction proposed by Jansen et al in 1997, we give out in this paper a generalized Dikin-type affine scaling algorithm for solving the P-*(kappa)-matrix linear complementarity prob...Based on the generalized Dikin-type direction proposed by Jansen et al in 1997, we give out in this paper a generalized Dikin-type affine scaling algorithm for solving the P-*(kappa)-matrix linear complementarity problem (LCP). Form using high-order correctors technique and rank-one updating, the iteration complexity and the total computational turn out asymptotically O((kappa + 1)root nL) and O((kappa + 1)n(3)L) respectively.展开更多
文摘对P*(κ)线性互补问题提出了一种自适应全-Newton步不可行内点算法.算法是对Mansouri等人(H.Mansouri and M.Pirhaji in Journal of Operations Research Society of China 1:523-536,2013)提出的单调线性互补问题的自适应不可行内点算法的推广.在算法的每一次迭代中,障碍校正参数θ的取值并不固定,它总在1/(51n(1+4κ)2)和1/(14n(1+4κ)2)之间取满足算法要求的最大值,使得算法快速收敛于问题的一个ε-近似解.
文摘Based on the generalized Dikin-type direction proposed by Jansen et al in 1997, we give out in this paper a generalized Dikin-type affine scaling algorithm for solving the P-*(kappa)-matrix linear complementarity problem (LCP). Form using high-order correctors technique and rank-one updating, the iteration complexity and the total computational turn out asymptotically O((kappa + 1)root nL) and O((kappa + 1)n(3)L) respectively.