Research On Train Operation Control Optimization Algorithm

Railway transportation has become one of the largest energy consumption departments in national economy, and train operation occupies a significant proportion of the whole energy consumption in railway system. In order to reduce the energy cost in train operation process, scholars throughout the world have carried out extensive researches on energy-saving based train operation optimization. However, train running process belongs to a multi-objective optimization problem, which needs a comprehensive consideration of factors including energy cost, operation time and stopping accuracy, besides, the optimal solutions of such optimization problem never exist in an absolute sense. Therefore, the research on how to solve train multi-objective optimization problem and choose optimal solutions according to different train operating conditions has profound theoretical and practical significance.The paper’s research emphasized on following three aspects:(1) An in-depth study of train operation optimization theory is developed, including train kinetic theory, train operating strategies and conditions choosing principles; train multi-particle dynamic model is formulated and train running process calculation method is analyzed in detail; train energy consumption, operation time and stopping accuracy optimization models are established based on train dynamic model constraints; on the basis of proving train operation optimal solutions’existence, the basic form of train operation optimal solutions is studied, and train operation optimal solutions’solving method is proposed.(2) Standard differential evolutionary algorithm (DE) and simulated annealing algorithm (SA) are carefully studied and compared on the aspects of advantage and disadvantage in solving multi-objective optimization problem, the possibilities of these two algorithms’combination are verified, and a hybrid evolutionary algorithm (HEA) is proposed from DE and SA; combining train operation optimization models and multi-objective optimization theory, all these three algorithms are applied to train operation Pareto optimal solutions’solving simulation, and the simulation results show that HEA has the fastest computing speed, most uniform Pareto optimal solutions’ distribution and widest Pareto fronts’distribution range, which proves HEA has the best performance in solving train optimization problem.(3) Based on Beijing-Shanghai high-speed railway datas, a careful simulation and analysis on HEA’s application in train operation optimization is conducted. Train single and multiple objective optimization processes are firstly simulated and studied, which proves train operation optimization belongs to a multi-objective optimization problem; secondly, simulation cases of different operating lines are designed, and train running optimization processes in these operating lines are simulated, the crucial factors which influence the optimal solutions’diversity are analyzed, and decision-making principles of choosing train optimal solutions under different circumstances are summarized.