改进马尔可夫链的船舶分段制造车间排产优化OA

The improved Markov chain is introduced to study the scheduling optimization problem of ship subsection manufactur-ing workshop.Minimizing the longest manufacturing time of segmented workpiece is selected as the optimization objective.The heterogeneous disjunctive graph of workshop state is constructed,and the process and equipment state features of workshop are extracted from it by using graph neural network model.Taking the extracted state features as the input,a large-scale model is constructed,and the production scheduling optimization scheme is obtained by cooperating with each process and equipment agent through the model.The optimization problem of workshop scheduling is transformed into Markov decision process.By set-ting the state,action and reward function,the scheduling problem is modeled as a sequential decision problem.In this frame-work,the cooperative agent deep reinforcement learning model learns the optimal strategy by interacting with the environment.The reward function is used to evaluate the pros and cons of each decision,guide the agent to optimize towards the goal of mini-mizing the maximum manufacturing time,so as to realize the effective evaluation and selection of scheduling schemes.The ex-perimental results show that under the optimization scheme of this method,the minimum value of the longest manufacturing time of the processed parts of the frame,outer plate and main deck is 400.05 min,490.52 min and 601.33 min,respectively.The se-lected performance deviation ratio index value is less than 20,and the equipment does not have idle time gap,which improves the manufacturing efficiency of ship building sections.