吴靳, 戴明强, 王俊杰, 等. 基于学徒制算法的航母舰载机保障作业调度[J]. 中国舰船研究, 2022, 17(4): 145–154. doi: 10.19693/j.issn.1673-3185.02198
引用本文: 吴靳, 戴明强, 王俊杰, 等. 基于学徒制算法的航母舰载机保障作业调度[J]. 中国舰船研究, 2022, 17(4): 145–154. doi: 10.19693/j.issn.1673-3185.02198
WU J, DAI M Q, WANG J J, et al. Carrier-based aircraft operation support scheduling based on apprenticeship learning agorithm [J]. Chinese Journal of Ship Research, 2022, 17(4): 145–154. doi: 10.19693/j.issn.1673-3185.02198
Citation: WU J, DAI M Q, WANG J J, et al. Carrier-based aircraft operation support scheduling based on apprenticeship learning agorithm [J]. Chinese Journal of Ship Research, 2022, 17(4): 145–154. doi: 10.19693/j.issn.1673-3185.02198

基于学徒制算法的航母舰载机保障作业调度

Carrier-based aircraft operation support scheduling based on apprenticeship learning agorithm

  • 摘要:
      目的  针对舰载机的保障调度问题,提出一种基于学徒制的调度优化算法,以快速制定舰载机保障调度计划。
      方法  运用学徒学习思想,将专家示例中的已执行任务与未执行任务成对予以比较,构造样本集,训练基于航母甲板特征的保障任务排班分类器,然后以此为基础设计学徒制舰载机机群保障任务调度算法,并从求解结果、求解时间和资源分配方面与传统的遗传算法进行对比。
      结果  结果显示,采用学徒制算法制定的舰载机机群保障调度计划与传统的遗传算法相比结果相当,但收敛速度提升了近4倍,且相比遗传算法能更加平均地分配保障资源。
      结论  采用学徒制算法可以充分学习专家的经验,解决静态单目标的舰载机保障调度问题,能为研究动态多目标的舰载机保障调度提供基础。

     

    Abstract:
      Objectives  Aiming at the operation support scheduling of carrier-based aircraft, this paper proposes a scheduling optimization algorithm based on apprenticeship learning which can quickly generate a operation support schedule for a carrier-based aircraft fleet.
      Methods  Using the apprenticeship learning method, the executed and unexecuted tasks in expert demonstrations are compared in pairs to construct a sample set, and the support task scheduling classifier is trained based on the deck features of aircraft carrier. On this basis, a support task apprenticeship learning algorithm for a carrier-based aircraft fleet is designed and compared with the traditional genetic algorithm (GA) in terms of solving solution, solving time and resource allocation.
      Results  The results show that the operation support schedule obtained by the apprenticeship scheduling algorithm is equivalent to that by the traditional GA, but the rate of convergence is increased nearly fourfold, and the support resources are more evenly distributed.
      Conclusions  The apprenticeship scheduling algorithm proposed in this paper can adequately learn from expert experiences and solve the problem of static single-objective carrier-based aircraft support scheduling. As such, this study provides references for further research in the field of dynamic multi-objective carrier-based aircraft support scheduling.

     

/

返回文章
返回