The core concept for the multidimensional knapsack problem

摘要: We present the newly developed core concept for Multidimensional Knapsack Problem (MKP) which is an extension of classical one-dimensional case. The multidimensional problem defined in dependence a chosen efficiency function items, since no single obvious measure available MKP. An empirical study on cores widely-used benchmark instances presented, as well experiments with different approximate sizes. Furthermore we describe memetic algorithm and relaxation guided variable neighborhood search MKP, are applied to original problems. experimental results show that given fixed run-time, metaheuristics general purpose integer linear programming solver yield better solution when problems size.