Resource Optimization Of Storage Yard At Automated Container Terminal Based On Digital Twin

The production of the container terminal mainly revolves around the operations of container loading and unloading,container collection and container pickup.Most activities,such as loading and unloading,and circulation of containers take place in the storage yard.Faced with the huge workload of handover and storage tasks,there is also a variety of operating resources,which forces the storage yard to maximize the utilization of space and resource.It is the key to improve the operation efficiency of the storage yard.The business of the automated terminal yard mainly focuses on the collection and delivery of containers from external trucks and ships,which involves the scheduling and optimization of resources such as storage space,loading and unloading equipment,and transportation facilities.Taken yard resources of the automated container terminal as the research object,this paper proposes the optimization problem of multiple yard resources in different scenarios with the combination of digital twin and operation optimization.In the complex and changeable production environment,the allocation of space resources,the scheduling of handling resources,and the path planning of transportation resources are studied.Through the application of digital twin and resource optimization methods,the efficient planning of multi-task and multi-resources of automated container terminals can be realized.The main research contents of this paper are as follows:(1)Since the uncertain factors in the operation of the automated container terminal will affect the efficiency of handling containers,it is necessary to implement a flexible scheduling scheme that matches the time-varying environment.In this paper,an automated yard scheduling framework based on digital twin is proposed to monitor the disturbed scene during actual operation and visualize real-time data in virtual space to adapt to the time-varying environment.It is a great way to provide decision support for resource optimization at the storage yard.In the proposed framework,a virtual model of an automated yard is established,and a virtual space synchronized with the physical space is formed through data acquisition and transmission.Based on real-time information,a mathematical model is established to deal with the resource optimization problem in uncertain scenarios,and the algorithm is used to solve the problem.The scheduling scheme is input into the virtual yard,and a feasible solution can be obtained after optimization through simulation.The framework can be used in the scheduling systems,such as storage area allocation,yard crane scheduling,and AGV scheduling.Through the application of digital twin,it can obtain dynamic information in the production process from the physical yard in time and propose resource optimization strategies that match the actual operation scenarios.The framework is also the basis for developing a digital twin scheduling system of automated container terminals.(2)When planning yard container area resources,it is necessary to consider the volume of container tasks at the storage yard.As the quantity of containers that are about to enter the yard is unknown,this paper proposes a prediction method based on long short-term memory recurrent neural network.Through the processing and training of the historical data,a better container volumes prediction model is obtained.Under the storage allocation mechanism based on digital twin,the balance strategy is developed to address the problem of unbalanced allocation at the storage area.The task information of import and export containers can be obtained in real-time.On this basis,the storage area is allocated for the import and export containers transported by external container trucks under the two operations of container being collected to port and being picked from port.To minimize the time of handling and transporting all containers,that is,to reduce the presence time of trucks as much as possible,a mathematical programming model is established and a strategy for balancing the task volumes between container blocks is designed.Then,an adaptive genetic algorithm is proposed to solve this problem.By reducing the variance of container tasks handled in the container blocks,the goal of balancing the task distribution in different container blocks is achieved,and the concentrated operations and uneven use of equipment in the storage yard are avoided.(3)According to the randomness of the arrival time of external container trucks,the dynamic optimization of the yard crane is studied.In the actual operation process,the delay or early entry of external container trucks will interfere with the operation plan of yard cranes,resulting in unnecessary costs,such as the emptying of yard cranes and the waiting of external container trucks.Under the yard crane scheduling mechanism based on digital twin,,arrival information of external container trucks and the task types of containers handled by yard cranes can be obtained in real-time,the principle of first-come and first-served is used to arrange the loading and unloading operations of yard cranes with priority.Then,a strategy for yard cranes to dynamically match tasks of external container trucks is proposed.On this basis,considering the safety factors during the operation of yard cranes,a multi-objective model with the shortest operating time of yard cranes and the waiting time of external container trucks and yard cranes is constructed,and a non-dominated sorting genetic algorithm Ⅱ based on heuristic rules is designed as the solution algorithm.A reasonable yard crane scheduling plan will be obtained for the operation at the storage yard.(4)Aiming at the scheduling problem of AGVs for horizontal transportation equipment in the scenario of unbalanced supply and demand of containers,the storage yard and the quay crane are set as the supply side and the demand point respectively.Under the AGV scheduling mechanism based on digital twin,information on AGV transportation tasks and route networks is obtained.By planning transportation tasks and transportation paths for AGV,the supply and demand between the quay crane and the storage yard can achieve a dynamic balance.With the goal of minimizing the delay time of task completion,a task allocation model of AGV is established.In addition,with the goal of minimizing the congestion rate,a path planning model for AGV is established.The ratio of the waiting time caused by conflicts and total transportation is used to measure the congestion of the AGV transportation path.At the same time,according to the idea of the time road network model,the purpose of balancing the distribution of traffic load is achieved by limiting the number of AGVs passing through the path.As a result,the possibility of congestion or collision in the same road section can be reduced.Then,the idea of the contract network protocol in the multi-agent system and machine learning is introduced to design algorithms which can solve task assignment and path planning of multi-AGV.(5)On the basis of the automatic yard scheduling framework based on digital twin,the application of the digital twin scheduling system at the automated container terminal is proposed.According to the design of components in the scheduling framework,such as the data center and virtual yard,the automated terminal digital twin scheduling system can be involved in data input,resource scheduling,simulation optimization and result visualization modules.With the background of a certain automated container terminal,taking the multi-resource optimization in different scenarios proposed as examples,the information of related resource tasks is imported into the data input module.Then,the corresponding model and solution algorithm are called in the resource scheduling module,and the obtained scheduling scheme is input into the simulation optimization module.According to the key indicators of the operation efficiency at automated terminals,the evaluation indicators of different resource optimization schemes are proposed,which are used to verify the validity and applicability of the AGV planning schemes.In this paper,an automated yard scheduling framework based on digital twin is established,and the multi-resource optimization of storage yard is proposed to be solved in different scenarios.The scheduling system based on digital twin at automated container terminals is applied to verify the feasibility of the method described in this paper.The real-time information obtained from the actual automated container terminal by digital twin can help the optimization of resource scheduling.It will provide the decision support for operation management at automated container terminals.In this paper,an automated yard scheduling framework based on digital twin is built to solve the resource optimization problem of automated container terminal yard in different scenarios.Aiming at the problem of container space allocation when the number of containers in the yard is uncertain,a container allocation mechanism based on digital twin is proposed,a mixed integer programming model with the least operating time is established,and an adaptive genetic algorithm based on container allocation rules is designed.For the scheduling problem of yard cranes when the external container trucks arrive randomly,the scheduling mechanism of yard cranes based on digital twin is proposed.Then,a mathematical model is established to minimize the operation time and waiting time of equipment,and a Non-dominated Sorting Genetic Algorithm Ⅱ based on dynamic rules is designed to solve the problem.For the multi-AGV scheduling problem in the yard when the supply and demand of container tasks are unbalanced,an AGV scheduling mechanism based on digital twin is proposed.A two-stage model with the minimum task delay time and path congestion rate is established,and a multi-agent Q-learning algorithm based on contract network protocol is designed.In the automated storage yard digital twin scheduling system,taking an automated container terminal as an example,the feasibility of the method proposed in this paper is verified.Hence,the optimization of the storage yard resource scheduling scheme is realized,which can provide guidance for the operation decision at the automated container terminal.