ew discuss several aspects of the parallel and the VLSIimplementa- tion of the interior point algorithms for solving thelinear program- ming problem. Various architectures, based on thedifferent parti- tioning of the input data, are proposed and thecomplexity of these implementations are discussed. Am economicinterpretation of a par- titioning scheme is then presented whichmight prove useful for the further improvements in the efficientparallelization of the interior point methods.
展开▼
