An improved binary search algorithm for the Multiple-Choice Knapsack Problem

Abstract

The Multiple-Choice Knapsack Problem is defined as a 0-1 Knapsack Problem with additional disjoint multiple-choice constraints. Gens and Levner presented for this problem an approximate binary search algorithm with a worst case ratio of 5. We present an improved approximate binary search algorithm with a ratio of 3 + (1/2)(t) and a running time O(n(t + log m)), where n is the number of items, m the number of classes, and t a positive integer. We then extend our algorithm to make it also applicable to the Multiple-Choice Multidimensional Knapsack Problem with dimension d.