Research On Vehicle Scheduling Optimization Of Quay Crane Assembly Process For Dynamic Tasks

The quay crane is the key handling facilities in the apron,and its assembly process accounts for more than 50% of the manufacturing cycle,which is of great significance for product delivery.Due to its complicated structure,the assembly process usually needs to coordinate the material transportation of dozens of parts processing workshop.Higher transportation requirements are put forward for vehicle scheduling due to frequent,diverse and dynamic transportation tasks.In the final assembly,the delay of the material may affect the completeness of the assembly process,and the impact may be amplified and the entire product will not be delivered on schedule in long assembly process.This paper studies the vehicle assignment in dynamic environment and the vehicle path planning under real-time road conditions.To improve the material transportation efficiency in the quay crane assembly process through vehicle scheduling optimization so as to endure the product to be delivered on schedule.The work is as followed:(1)The vehicle assignment method in the dynamic environment is studied in this paper in order to ensure that the vehicle can effectively respond to the transportation task.A distributed reinforcement learning algorithm which vehicle is the decision agent is designed with the vehicle being a transportation bottleneck.Based on the business logic of the vehicle assignment,the Markov model is established and its corresponding state,action and reward are designed accordingly.In view of the fact that complex environment including many influencing factors in the vehicle assignment process and the dynamic environment,the algorithm is optimized based on the traditional Q-learning algorithm,adopting the neural network-based value function approximation method and the convergence mechanism based on empirical playback.Finally,the vehicle assignment optimization in the dynamic environment is realized by the problem-oriented reinforcement learning algorithm design.(2)The vehicle path optimization method under real-time road conditions is studied in this paper in order to ensure the efficient vehicle transportation.The actual road caondition is simulated based on Plant Simulation,and the congestion degree is extracted to describe the rout condition.Based on that,the mathematical model with the optimization goal of the shortest transportation time is established and an improved particle swarm optimization algorithm is designed.The proposed algorithm is validated and analyzed by simulating different problem scenarios.Finally,combined with the simulation model to describe the real-time road conditions,the timeliness of the path optimization algorithm is improved,and the real-time optimization of the vehicle path is realized.(3)The simulation model is to built based on Plant Simulation to verify the method above based on the background of an assembly enterprise.The results shows that the proposed method improves the current resource scheduling waste to some extent caused by the scheduling rules such as “Nearest Neighbor” and “first-come first-served service”,and realizes the intelligent scheduling decision.In summary,the research results of this paper can meet the needs of vehicle scheduling in the actual assembly process of quay crane,which is of great significance for improving the material transportation efficiency required for final assembly.