Two YC Dynamic Scheduling Q Learning Algorithm For Container Terminal

Economic globalization has made the world more and more closely connected and the trade exchanges more and more frequent.The rapid development of information technology has accelerated the process of globalization.The advantage of container and convenient condition make container transportation more and more prominent in the position of goods transportation.As a transit station of container transportation,container terminal is faced with opportunities and challenges.The growth of container transportation directly leads to the increase of container throughput.With the development of upsizing of container ships,the competition among ports is becoming fiercer.To stand out from the competition,container terminals need to improve their core competitiveness.In addition to the port’s own geographical location and hinterland economy,container terminals need to improve their production technology,service and management level.Container yard is the place for temporary storage of containers in the container terminal.It is the intermediate link of the operation chain of container terminal and plays an important role in the operation of the whole terminal.As an important part of container terminal yard management,the optimization of container terminal yard and YC scheduling can not only improve the working efficiency of the yard,but also relieve the working pressure of the wharf and improve the working efficiency of the wharf,so as to improve the competitiveness of the wharf.In this paper,the scheduling problem of two YC in container terminal yard under dynamic environment is studied.Based on the analysis of container terminal yard operation system,this paper analyzes the operation flow of container terminal and the content of yard YC scheduling,and then analyzes the influencing factors of YC scheduling of container terminal.According to the dynamic environment conditions of YC scheduling problem,the Markov decision process model of YC scheduling under the dynamic environment is established.Aiming at the Markov decision process model of two YC scheduling for container terminals in a dynamic environment,a q-learning algorithm for two YC scheduling for container terminals is designed,and the algorithm exploration strategy is improved.Finally,four basic YC scheduling schemes and nine two YC scheduling schemes based on Q learning algorithm are designed for performance experimental analysis.The results show that: 1)considering only the current task of the yard,the Q learning algorithm can effectively schedule YC job and reduce the average waiting time of truck by 0.5%~20.8%.2)compared with two YC scheduling Q learning algorithm that only considers the current task of the yard,two YC scheduling Q learning algorithm using the task information of the yard in the future period can reduce the average waiting time of truck by more than 15%.3)in the use of storage area in the future for a period of time task information two YC scheduling Q learning algorithm on the basis of adding yard busy is swept off YC,and using the yard in the future for a period of time the task of information compared to two YC scheduling Q learning algorithm,effectively reduced in the yard is busy fixation of the average waiting time of truck.