共享适应度粒子群在双机ETV中的应用

The application of Shared fitness particle swarm in double ETV system

摘要:

针对机场货运区双机双货位升降式转运车（ETV）并行工作时任务链生成困难的问题，提出一种基于ETV载物台的任务链生成算法。该算法结构清晰，且易于计算总运行时间。针对粒子群算法在计算双机ETV最优任务序列时易出现早熟问题，提出改进的共享适应度粒子群算法。该算法在混沌粒子群算法的基础上融合共享适应度的思想，当混沌粒子群算法的全局最优解趋于稳定时，选取百分之二十的粒子留守共享半径内，其他粒子重新初始化并继续迭代。实验仿真结果表明和标准粒子群算法、混沌优化粒子群算法相比，共享适应度粒子群算法可以有效地避免早熟，全局寻优能力更强，得到的结果更优且更稳定。

Abstract:

Aiming at the difficulty of task chain generation when two EVTs work in parallel in the airport cargo area, a task chain generation algorithm based on ETV platform is proposed. This algorithm has a clear structure and easy to calculate the total running time. Aiming at the problem of premature maturity of particle swarm optimization algorithm, an improved shared fitness particle swarm optimization algorithm is proposed. Twenty percent of the particles are selected to be in the left-behind Shared radius, and other particles are re-initialized and iterated, when the global optimal solution of the algorithm tends to be stable. The simulation shows that, compared with the standard particle swarm algorithm and chaotic particle swarm algorithm, the Shared fitness particle swarm algorithm can effectively avoid the precocity, and the global optimization ability is stronger. The solution is better and more stable.