Research On Emergency Port Evacuation Decision-making In Intelligent VTS

The safety and efficiency of ship entry and exit operation is a common concern for shipping enterprises and maritime functional departments.The emergency port evacuation decision is to study the evacuation problem of ships under the port congestion.In view of the nonlinear and uncertain characteristics of ship movement,this thesis studies the emergency evacuation decision problem under deterministic and fuzzy conditions,aiming to provide new solutions for VTS in ship risk management and ship evacuation,and further improve the traffic efficiency of port.Firstly,a ship static data quality evaluation system is constructed by using data correlation analysis and processing techniques and detecting outliers with the help of characterization learning algorithm.Secondly,combining the cleaned trajectory data and emergency port evacuation dynamic data with collision detection methods in graphics,a ship trajectory conflict detection framework was established and applied to regional risk assessment.Then,combining mathematical planning method,fuzzy optimization method,graph theory and intelligent optimization algorithm,corresponding decision models are established and optimization solution algorithms are designed for different emergency port evacuation decision problems.The main research work of this thesis is shown as follows.(1)Based on the representation recognition method,the abnormal state in the static information of the ship is mined and the abnormal detection model of the static information of the ship is established.The ship trajectory equal interval discretization method is improved,the variable time trajectory segmentation method to realize the spatio-temporal alignment of ship trajectory is proposed,and the ship trajectory conflict detection algorithm is proposed based on mentioned above.The ship trajectory conflict detection framework is established based on SAT algorithm and ship domain,and realized the application in the assessment of ship collision risk at sea.(2)An emergency port evacuation decision model with deterministic conditions under one-way channel is established based on the nonlinear ship motion.The linear transformation method of the nonlinear decision model is proposed by introducing the concept of MSTI.The MSTI calculation method is determined by combining with the ship trajectory conflict detection algorithm.The optimal solution method of the decision model is designed based on genetic algorithm.The model and algorithm are validated by combining numerical simulation using CPLEX solver and priority method,and the results show that the optimization algorithm can obtain the value of the approximate exact solution with an error of about 6.83%.(3)The emergency port evacuation decision model with fuzzy conditions under one-way channel is established based on fuzzy optimization method.The linear transformation of the tidal height nonlinear constraint problem is realized by introducing the concept of navigable time window,and the algorithm for calculating the navigable time window of the ship is determined.Based on the fuzzy rules and operation methods,the genetic algorithm evaluation function and operation operator are improved to realize the optimal solution under fuzzy conditions,and the sensitivity of the parameters of the optimization algorithm is analyzed.The model and algorithm are validated by combining numerical simulation and priority methods,and the results show that the Makespan performance is improved by 23% and 30% in terms of lower and upper limits of fuzzy numbers,respectively.(4)An emergency port evacuation decision model based on undirected graph for deterministic conditions under waterway network.Based on the traffic conflict technique,the vessel traffic conflicts within the waterway network are summarized and classified,and the method of eliminating the vessel traffic conflicts within the waterway intersections and segments is proposed.Based on the MSTI concept,the STI concept applicable to the channel network is proposed and the calculation process and method are given.The model optimization solution method is designed by combining the knowledge of graph theory and genetic algorithm,and the model and algorithm are validated for restricted and Y-shaped channels,respectively.The results show that the performance of Makespan is improved by 64.5% in the restricted channel and 60.3% in the Y-shaped channel compared with the priority rule-based method.