[1]WANG Guo-lei,ZHONG Shi-sheng,LIN Lin.Bilevel Qlearning algorithm for dynamic multimachinescheduling problems[J].CAAI Transactions on Intelligent Systems,2009,4(3):239-244.
Copy

Bilevel Qlearning algorithm for dynamic multimachinescheduling problems

CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume: 4 Number of periods: 2009 3 Page number: 239-244 Column: 学术论文—智能系统 Public date: 2009-06-25
Title:
Bilevel Qlearning algorithm for dynamic multimachinescheduling problems
Author(s):
WANG Guo-lei;  ZHONG Shi-sheng;  LIN Lin
School of Mechanical Engineering, Harbin Institute of Technology, Harbin 150001, China
Keywords:
dynamic multimachine scheduling;  Qlearning;  action set;  state space division;  reward function
CLC:
TP273
DOI:
-
Abstract:
Traditional Qlearning is very effective in dynamic singlemachine scheduling problems, yet sometimes it cannot get optimal results for dynamic multimachine scheduling problems due to its lack of global vision. To resolve this, a twolayer Qlearning algorithm was put forward. The bottomlevel of Qlearning was focused on localized targets in order to learn the optimal scheduling policy which can minimize machine idleness and the mean flow time of single machines. On the other hand, the toplevel of Qlearning was focused on global targets in order to find the dispatching policy which can balance machine loads and minimize the overall tardiness of all jobs. The scheduling and dispatching rules of agents, the method for dividing state space and the reward functions were all examined. Simulation results showed that the proposed twolayer Qlearning algorithm can improve the results of dynamic multimachine scheduling problems.
References:
［1］严浙平, 李 锋, 黄宇峰. 多智能体Q学习在多AUV协调中的应用研究［J］. 应用科技, 2008, 35(1): 5760.
 YAN Zheping, LI Feng, HUANG Yufeng. Research on application of multiagent Qlearnling in multiAUV coordination［J］. Applied Science and Technology, 2008, 35(1): 5760.
 ［2］潘燕春, 冯允成, 周 泓,等. 强化学习和仿真相结合的车间作业排序系统［J］. 控制与决策, 2007, 22(6): 675679.
PAN Yanchun, FENG Yuncheng, ZHOU Hong, et al. Reinforcement learning integrated with simulation for jobshop scheduling system［J］. Control and Decision, 2007, 22(6): 675679.
［3］AYDIN M E,〖AKO¨〗ZTEMEL E. Dynamic jobshop scheduling using reinforcement learning agents［J］. Robotics and Autonomous Systems, 2000, 33(2/3): 169178.
［4］WANG Y C, USHER J M. Application of reinforcement learning for agentbased production scheduling［J］. Engineering Applications of Artificial Intelligence, 2005, 18(1): 7382.
 ［5］WANG Y C, USHER J M. Learning policies for single machine job dispatching［J］. Robotics and Computer Integrated Manufacturing, 2004, 20(6): 553562.
［6］魏英姿,赵明扬. 强化学习算法中启发式回报函数的设计及其收敛性分析［J］. 计算机科学, 2005, 32(3):190193.
WEI Yingzi, ZHAO Mingyang. Design and convergence analysis of a heuristic reward function for reinforcement learning algorithms［J］. Computer Science, 2005, 32(3): 190193.
［7］王世进,孙 晟,周炳海,等. 基于Q学习的动态单机调度［J］. 上海交通大学学报, 2007, 41(8): 12271232.
 WANG Shijin, SUN Sheng, ZHOU Binghai, et al. Qlearning based dynamic single machine scheduling［J］. Journal of Shanghai Jiaotong University, 2007, 41(8):12271232.
［8］杨宏兵,严洪森. 知识化制造系统中动态调度的自适应策略研究［J］. 控制与决策, 2007, 22(12): 13351340.
YANG Hongbing, YAN Hongsen. Adaptive strategy of dynamic scheduling in knowledgeable manufacturing system［J］. Control and Decision, 2007, 22(12): 13351340.
［9］WATKINS C, DAYAN P. Technical note: Qlearning［J］. Machine Learning, 1992, 8(3/4): 279292.
Similar References:

Memo

-

Last Update: 2009-08-31

Copyright © CAAI Transactions on Intelligent Systems