FAILED
首站-论文投稿智能助手
典型文献
Manufacturing Resource Scheduling Based on Deep Q-Network
文献摘要:
To optimize machine allocation and task dispatching in smart manufacturing factories,this paper proposes a manufac-turing resource scheduling framework based on reinforcement learning(RL).The framework formulates the entire scheduling process as a multi-stage sequential decision problem,and further obtains the scheduling order by the combination of deep convolu-tional neural network(CNN)and improved deep Q-network(DQN).Specifiically,with respect to the representation of the Mar-kov decision process(MDP),the feature matrix is considered as the state space and a set of heuristic dispatching rules are denoted as the action space.In addition,the deep CNN is employed to ap-proximate the state-action values,and the double dueling deep Q-network with prioritized experience replay and noisy network(D3QPN2)is adopted to determine the appropriate action accord-ing to the current state.In the experiments,compared with the tra-ditional heuristic method,the proposed method is able to learn high-quality scheduling policy and achieve shorter makespan on the standard public datasets.
文献关键词:
作者姓名:
ZHANG Yufei;ZOU Yuanhao;ZHAO Xiaodong
作者机构:
School of Electronic and Information Engineering,Tongji University,Shanghai 201804,China
引用格式:
[1]ZHANG Yufei;ZOU Yuanhao;ZHAO Xiaodong-.Manufacturing Resource Scheduling Based on Deep Q-Network)[J].武汉大学自然科学学报(英文版),2022(06):531-538
A类:
Specifiically,D3QPN2
B类:
Manufacturing,Resource,Scheduling,Based,Deep,Network,To,optimize,machine,allocation,task,dispatching,smart,manufacturing,factories,this,paper,proposes,resource,scheduling,framework,reinforcement,learning,RL,formulates,entire,process,multi,stage,sequential,decision,problem,further,obtains,order,by,combination,deep,convolu,neural,network,improved,DQN,respect,representation,Mar,kov,MDP,feature,matrix,considered,state,space,heuristic,rules,denoted,action,In,addition,employed,proximate,values,double,dueling,prioritized,experience,replay,noisy,adopted,determine,appropriate,accord,current,experiments,compared,tra,ditional,method,proposed,able,high,quality,policy,achieve,shorter,makespan,standard,public,datasets
AB值:
0.615465
相似文献
机标中图分类号,由域田数据科技根据网络公开资料自动分析生成,仅供学习研究参考。