作者: Eli Hellerman , Dennis C. Rarick
DOI: 10.1007/978-1-4615-8675-3_6
关键词: Linear programming 、 Prime (order theory) 、 Basis (linear algebra) 、 Transformation (function) 、 Set (abstract data type) 、 Product (mathematics) 、 Computation 、 Algorithm 、 Computer science 、 Point (geometry)
摘要: Production linear programming systems often use the product form of inverse (PFI) to effect necessary computations. Experience has shown that PFI or techniques closely related are among more efficient computational methods available in terms speed, storage utilization, accuracy, control round-off errors and general flexibility. However, a characteristic is an additional amount work performed each successive iteration because new transformation vector (η ) added with pivot. The η, as well those previously generated, processed both forward backward on succeeding iteration. Therefore, economics this situation demands at some point current basis be reinverted expressed compact concise set vectors (ETA). Both speed preservation sparsity prime considerations during reinversion no progress made towards optimization process sparse ETA will interations. To these ends we have developed “Partitioned Pivot Procedure” which refer “P4”.