Ship Weather Routing Optimization Based On Evolutionary Algorithm

With the frequent world trade activities,ship transportation has been continuously developed.During an ocean-going voyage,ships are faced with many problems,such as long sailing time,high burnup cost and unstable meteorological conditions.Therefore,the research of ship weather routing based on evolutionary algorithm is carried out.First,the waypoint adjustment algorithm and the waypoint generation algorithm are presented to solve the actual constraint problem.When we use the waypoint to represent a route,the evolutionary algorithm can make the route design have very high degree of freedom,but it also brings some hidden dangers.The waypoint may appear in an unreasonable position by the crossover and mutation operation,so,the waypoint adjustment algorithm is used to solve the mismatch between waypoint’s order and position,and the route point generation algorithm is designed for the case where the waypoint falls into the forbidden area.Second,dynamic weather routing scheme is given.Ocean-going routes span a wide range of time and space,and the marine meteorological environment has great differences in time and space,which cause that the meteorological changes should be fully taken into account in the estimation of voyage time and burnup.In this paper,the dynamic programming method is used to realize the real-time updating of the arrival time and cumulative burnup of the remote route point,meanwhile,in order to reduce the estimation error caused by adjacent waypoints far away,the marine route segment is refined to improves the real-time and accuracy of route time and burnup estimation according to the time accuracy of meteorological data.Third,the optimization algorithm of weather routing in uncertain variable space is given in this paper.For route planning,the number of waypoints is usually preset to reduce the complexity of the problem,and the experimental results show that the optimization results of the shortest voyage time or the least burnup route are differentunder different waypoint number,which means that when we optimize the route with the evolution algorithm,the change of the waypoint number should also be considered.In order to solve the problem that the length of the variables is inconsistent between the two sides of the crossover,this paper proposes additional waypoint algorithm and waypoint deletion algorithm based on the dynamic weather routing scheme,and on this basis,gives a structural adjustment algorithm to deal with the crossover of uncertain variables.Then,the influence of different numbers of waypoint on route optimization can be solved.Fourth,a multi-objective optimization algorithm is presented for weather routing.The simulation results show that the voyage time and burnup are in conflict,so the optimization of weather routing can also be regarded as a multi-objective optimization problem.According to the high dependence of ship routes on expected arrival time,an undominated adaptive MOEA/D algorithm is proposed in this paper.In this algorithm,the PBI method in MOEA/D is replaced by a non-dominant sorting method.Meanwhile,the weight vector can be updated and homogenized automatically in the iterative process,and its adaptability ensures the width and uniformity of the distribution.