Research On Operation Optimization Of Bulk Cargo Port Based On Condition Detection Data And Swarm Intelligence Algorithm

With the development of information and intelligence in recent years,"intelligent transportation" has become the main direction of modern transportation development.As the key node of the transportation system,the port promotes the construction of "port station integration" and improves port operation efficiency,which is an important practice of the "intelligent transportation" concept.Aiming at the problems of unreliable operation data and unscientific production scheduling in the production process of unloading and loading of bulk cargo terminals,this dissertation studies the port’s data collection,equipment management,monitoring data processing,unloading scheduling,berth scheduling,loading scheduling and throughput prediction.This dissertation systematically analyzes the data requirements of port status monitoring,and then design and develop a data integration system for modern bulk port status monitoring to realize automatic data collection,standardized definition,and perform error processing and data compression on the collected integrated data,which provides a data basis for port unloading scheduling,berth scheduling,loading scheduling and throughput prediction.In view of the intelligent optimization of unloading tasks in ports,the mathematical modeling and scheduling optimization algorithm of unloading operations are studied.Based on the relevant data of unloading operations in the port condition monitoring system,considering factors such as multiple cargo types,train scheduling plans,availability of unloading equipment and their mutual constraints,a task model of unloading operations is constructed.An improved salps optimization algorithm was proposed,which introduced adaptive inertia weight and random Cauchy mutation strategy.The algorithm is used to optimize the unloading task,so that the train’s residence time in the port is greatly reduced.For the intelligent optimization of port’s berth scheduling,the research on mathematical modeling and optimization algorithm of berth scheduling problem is carried out.Based on the relevant data of berth scheduling in the port status monitoring system,and considering factors such as berth status,number of ships,loading and unloading time,an optimization model for the shortest ship time in port is constructed.An improved gray wolf algorithm(IGWO)is proposed,which uses the Sin chaotic sequence for initialization,and introduces leading wolf strategy,cooperative competition mechanism,and adaptive weights to enhance the information exchange between individuals,improve the utilization of information and accelerate the convergence speed of the algorithm.The improved gray wolf algorithm is used to optimize the berth scheduling,which reduces the total dwell time of all ships and improves the efficiency of berth utilization.Aiming at the intelligent optimization of ship loading and scheduling,it is studied based on reinforcement learning method.Based on the actual port environment,the yard information,ship demand information and equipment information are extracted,and a Markov decision process model in line with the actual production situation is established;The description of state space and agent action space of port environment is designed.Based on the characteristics of Double DQN deep reinforcement learning algorithm and scheduling problem model,an improved network structure and action function are proposed and improved ε-greedy explores strategies.The training is based on the improved algorithm,and the simulation experiment is carried out by simulating the actual scheduling situation of the port.The optimal scheduling scheme is obtained under the conditions of random arrival of tasks and random state of storage yard.The improved production scheduling plan reduces the total working time and improves equipment utilization.Aiming at the problem of intelligent forecasting of annual throughput of bulk cargo ports,this paper studies the throughput forecasting model..The throughput-related variables are analyzed,and a back-propagation neural network based on ant colony algorithm is proposed to optimize the initial weights and thresholds of the BP neural network.Based on the condition monitoring data in the port,a forecast model of port throughput is established.The improved BP prediction model can effectively improve the prediction accuracy of the original BP,and obtain higher prediction accuracy in the case of a small number of samples.