AN EXACT ALGORITHM FOR LARGE UNBOUNDED KNAPSACK-PROBLEMS

Given n types, each having an associated profit and weight, and a container of given capacity, the unbounded knapsack problem is to determine the number of items of each type to be selected so that the corresponding total profit is a maximum and the corresponding total weight does not exceed the capacity. We present upper bounds, dominance relations, and an approach-based on the definition of a core problem-to the exact solution of very large instances of the problem. We give the results of computational experiments on randomly generated test problems involving up to 250 000 item types. © 1990.