Abstract: The Time-Varying Knapsack Problems（TVKP） is a classic dynamic combinational optimization problem. For solving TVKP based on PSO, a greedy Particle Swarm Optimization with Double-structure coding（DGPSO） is proposed, which uses greedy strategy to modify the void coding of individual. In order to verify the validity and efficiency of DGPSO, two instances of TVKP are used. The results show that the capability that DGPSO adjust to a new environment is more optimization, and it is better that PDGA scout for optimized solutions.

摘要： 时变背包问题（TVKP）是一种典型的动态组合优化问题，由于其中某些量的动态变化，导致此问题非常难以求解。基于双重结构编码微粒群算法（DPSO）与贪心修正策略（GCOS）相结合，给出了一种求解TVKP 的新方法，通过对2个大规模TVKP实例的仿真计算表明：该方法比原对偶遗传算法适应环境变化能力和跟踪最优解的能力更强，非常适于求解TVKP问题。