A Harmony Search Algorithm for Resolving Multi-Dimensional Knapsack Problems

A multi-band encoding harmony search algorithm is proposed for MKP in this work.First,harmonic memory is initialized by multi-band encoding.Adaptive harmony memory consideration rate,pitch adjusting rate and band width are used to produce new solution.The infeasible solution is modified by modified algorithm for satisfy the constraint condition.Then,a pareto-optimal-set optimization algorithm is used to enlarge coverage rate and uniformity.Finally,20 independent experiments illustrate that the proposed algorithm has some advantages in convergence velocity,population diversity and solution precision compared with traditionally algorithm.