Solving Large Scale Generalized Networks
John M. Mulvey, Stavros A. Zenios · Journal of Information and Optimization Sciences · 1985
The efficiency of a generalized network optimization program is investigated. The computational experiments are aimed at improving the internal tactical rules of the algorithm. Alternative pivot strategies are presented in conjunction with a dynamically-adjusted approach. Column normalization factors and the Big-M starting method are also analyzed. It is shown that carefully constructed internal tactics can improve the performance of the generalized network program, especially for large scale examples. The newly developed code compares quite favorably with a mature state-of-the-art program–NETG.