Research On Multi-object Route Planning Of Ships Under Hydrological And Meteorological Conditions

With the rapid development of the economy and shipping industry,people have put forward higher requirements for the safety,comfort,greenness,and economy of ship navigation.Ships are inevitably faced with complex external environments during navigation,such as narrow waterways,bridges,water depths,other ships in inland rivers,and islands and reefs,coastlines,buoys,and severe wind waves that seriously affect the ship’s navigational status during the long-distance voyage.In order to ensure the safety of ships,it is necessary to plan efficient routes for inland and ocean-going ships.Therefore,based on particle swarm and genetic algorithm,this paper designs a simulation system for ship route planning in inland river environment and ocean environment.The main research contents and results are as follows:(1)The mathematical model construction of ship route planning under hydrological and meteorological conditions is studied.First,based on the characteristics of particle swarm algorithm and genetic algorithm,a method for constructing a ship sailing area model is proposed,which limits the ship’s navigable area to the local sea area around the usual route,which reduces the amount of calculation and ensures accuracy.Then,a mathematical representation method of the ship route model is proposed.The route is composed of multiple route sections including heading,still water speed,and other parameters.Finally,the main constraints that need to be considered in route planning are discussed and analyzed;(2)The single-object and multi-object route planning algorithms of ships in inland river environments are studied.First,the method of constructing the inland river map model is analyzed,the map of the inland navigation area is decomposed into a raster map,and navigable and non-navigable areas are established in the raster map.Then a single-target route planning algorithm based on an improved genetic algorithm is proposed,which has good fast convergence and searchability for route solutions.Finally,an improved multi-objective route planning algorithm is proposed based on the non-dominated sorting genetic algorithm,which improves the quality of the route solution through a multi-group elite selection operation.(3)The single-target and multi-target route planning algorithms of ocean-going ships are studied.First,a multi-objective weather route planning framework is proposed.The framework consists of six parts,namely optimization criteria,ship speed analysis,model construction,multi-objective algorithm,route evaluation,and route selection.Secondly,a mathematical model of the objective function of the ship in the meteorological route optimization is established,which mainly includes sailing time,fuel consumption,sailing risk,and arrival at a fixed time.Thirdly,because the wind and wave conditions have a greater influence on the ship’s speed,the calculation method of the ship’s speed loss under wind and wave conditions is studied.Then,a single-target route planning algorithm based on an improved particle swarm algorithm is proposed,which introduces crossover and mutation operations,and improves the speed update formula.Finally,a hybrid particle swarm optimization genetic algorithm is proposed to solve the multi-objective route planning problem.The algorithm not only has the characteristics of fast convergence speed of particle swarm optimization but also improves the diversity of route solutions by combining crossover,mutation,and multigroup elite selection operation.In order to avoid the uneven distribution of the Pareto front solution,this paper proposes a method to improve the Pareto optimal front distribution.According to the recommended route selection criteria,the shortest flight time route,the least fuel consumption route,the lowest risk route,and the recommended route can be obtained in the Pareto optimal solution set.(4)A simulation system for ship multi-target route planning has been established.The established simulation system has a good visual interface,in which algorithm parameters,navigation task parameters,navigation area maps,route data,algorithm optimization results,etc.can all be set or viewed in the interface.Based on this simulation system,this article designs a number of simulation experiments in different scenarios for single-target and multi-objective route planning.Experimental results show that the algorithm proposed has good convergence and route planning capabilities.