Research On Methodology Of Train Operation Adjustment Based On Artificial Fish Swarm Algorithm

In the process of operation,the train will suffer from the interference inevitably,which will cause the deviation from the planned operation diagram.When train deviates from the planned operation diagram,the current solution is that the dispatcher relying on his own experience and technique adjusts the train operation.However,the adjusted result depending on dispatcher's own knowledge and experience not only brings tremendously physical intensity to dispatchers,but also causes low precision and efficiency of adjusted result.Recently,for this situation,researchers have used methods of operations research,computer simulation as well as artificial intelligence to research the train operation adjustment.The train operation adjustment with the characteristics of large scale and various constraints cannot be easily solved by the exact algorithm so that researchers mainly have used artificial intelligence algorithm to research such problems in the current studies.The swarm intelligent algorithm which is a branch of artificial intelligence algorithm is applied to solve the problem of train operation adjustment in this thesis.The research on train operation adjustment will be conducted mainly from the following aspects:Firstly,through the collection of literatures on the train operation adjustment over the years,the merits and demerits of every method applied to adjust train operation are analyzed and compared in this thesis,and the principles and measures of train operation adjustment are summarized.The result shows that it is more practical to use swarm intelligence algorithm to achieve the train operation adjustment.Besides,artificial intelligence algorithms used in train operation adjustment in recent years are compared in this thesis,mainly including genetic algorithm,particle swarm optimization and ant colony algorithm.Secondly,the artificial fish swarm algorithm and simulated annealing algorithm are introduced in this thesis,including their principles,merits and demerits as well as application;at the same time,the model about train operation adjustment is established,containing its objective function and constraints.The specific steps are shown as follows: firstly,taking one segment in Zhengzhou-Xi'an high-speed railway as a simulation example,the artificial fish swarm algorithm is used to achieve the train operation adjustment.Meanwhile,it is programmed and simulated under the simulation environment MATLAB2016 a.It can be found that the result obtained by the artificial fish swarm algorithm is more accurate than those by the immune ant colony algorithm and genetic algorithm,except that the simulation time of former is longer.Then,combining the global convergence ability of artificial fish swarm algorithm with the rapid convergence ability of simulated annealing algorithm,a hybrid fish swarm optimization algorithm which combines simulated annealing algorithm with artificial fish swarm algorithm is proposed and applied to train operation adjustment inthis thesis.The result shows that the accuracy obtained by the hybrid fish swarm optimization algorithm is identical with that by the artificial fish swarm algorithm under the common simulation environment,except that the simulation time of former is greatly shortened.