作者: Michel Vasquez , Yannick Vimont
DOI: 10.1016/J.EJOR.2004.01.024
关键词: Multidimensional analysis 、 Mathematical optimization 、 Tabu search 、 Constraint (information theory) 、 Continuous knapsack problem 、 Knapsack problem 、 Linear programming 、 Algorithm 、 Cutting-plane method 、 Mathematics 、 Heuristic (computer science)
摘要: Abstract Geometric Constraint and Cutting planes have been successfully used to solve the 0–1 multidimensional knapsack problem. Our algorithm combines Linear Programming with an efficient tabu search. It gives best results when compared other algorithms on benchmarks issued from OR-L ibrary . Embedding this in a variables fixing heuristic still improves our previous results. Furthermore difficult sub problems about 100 500 original ones could be generated. These small are always very hard solve.