A Linked List Data Structure for a Binary Knapsack Algorithm.

摘要

This paper describes a novel application for a linked list in Greenberg's and Hegerich's algorithm for the binary knapsack problem. This list accelerates the solution of the problem by providing an efficient method for solving linear programming relaxations of the problem. Computational testing on 400 randomly generated problems with up to 7,500 variables is reported. For these problems, average solution times increase linearly with problem size when the linked list is used. The effectiveness of various problem reduction schemes is also explored.