The results from the computat]onal testmg of an algorithm designed to solve large-scale multiplechoice knapsack problems are described. Data list structures, sorting techniques, and fathoming criteria are investigated. It is determmed that considerable storage space can be ehmlnated from the original implementation of the algorithm without affecting execution time Also the insertion of a heap sort m the algorithm does provide a substantial reduction in computing time when solving the larger problems. Theoretical results relating to the multiple-choice knapsack are presented.
Unfortunately, ACM prohibits us from displaying non-influential references for this paper.
To see the full reference list, please visit http://dl.acm.org/citation.cfm?id=357458.