A Modified Genetic Algorithm for Multi-Objective Optimization on Running Curve of Automatic Train Operation System Using Penalty Function Method
- 84 Downloads
The running curve optimization of Automatic Train Operation system usually takes into account running time, energy consumption and passenger comfort. In this paper, in order to provide more comprehensive optimization and accurate reference of running curve for Automatic Train Operation system, we adopted the multi-objective optimization strategy of genetic algorithm to optimize from five aspects: speeding (safety), parking accuracy, punctuality, energy consumption and comfort. In order to increase the convergence speed of genetic algorithm to the optimal solutions, we propose a modified genetic algorithm, which the penalty function method is added into the fitness objective function. The modified genetic algorithm optimization program is written by M language in MATLAB, and combined with a graphical user interface tool to design the optimization system. Its validity is verified by comparison between the tests based on three different interstation of Shanghai Metro Line 11. The results show that it is effective and practicability to use the designed system to optimize the running curve of Automatic Train Operation system.
KeywordsAutomatic train operation (ATO) Modified genetic algorithm (MGA) Multi-objective optimization Running curve Urban rail transit
This work is supported by the National “Twelfth Five-Year” Pillar program for Science & Technology – the Interoperability Comprehensive Evaluation Integrative Platform and Demonstration for Urban Rail Transit (No.2015BAG19B02).
- 2.Zheng, W., Xu, H.Z.: Modeling and safety analysis of maglev train over-speed protection based on stochastic petri nets. J. China Railway Soc. 31(4), 59–64 (2009)Google Scholar
- 4.Chen, D., Tang, T., Gao, C., Mu, R.: Research on the error estimation models and online learning algorithm for Train Station parking in urban rail transit. China Railway Sci. 31(6), 122–127 (2010) (in Chinese)Google Scholar
- 16.Li, J.Q.: Analysis of the Train’s traction energy consumption of shanghai metro line 11. Mechatronics. 19(6), 32–35 (2013) (in Chinese)Google Scholar
- 17.Holland, J.: Adaptation in Natural and Artificial Systems, University of Michigan Press, p. 1975. USA, Ann Arbor, MI (1975)Google Scholar
- 20.Kuri-Morales, A.F., Gutiérrez-García, J.: Penalty Function Methods for Constrained Optimization with Genetic Algorithms: A Statistical Analysis. In: Coello Coello, C.A., de Albornoz, A., Sucar, L.E., Battistutti, O.C. (eds.) MICAI 2002: Advances in Artificial Intelligence. MICAI 2002. Lecture Notes in Computer Science, vol 2313. Springer, Berlin (2002)Google Scholar
- 21.Y. J. Lei, S. W. Zhang (2014) MATLAB genetic algorithm toolbox and its application. Xidian University Press, 2014 (in Chinese)Google Scholar
- 23.Syswerda, G.: Simulated crossover in genetic algorithm. Found. Genet. Algorithms. 2, 239–255 (1993)Google Scholar
- 25.Chen, Y., Qian, C.Y., Xi, X.D.: Traction energy consumption test and analysis for shanghai metro AC 16 electromotive train. Urban Mass Transit. 19(9), 34–38 (2016) (in Chinese)Google Scholar
- 26.Xu, K., Wu, L., Yang, F.F.: Automatic train operation system in urban rail transit based on PSO-ICS algorithm optimization. J. Railway Sci. Eng. 14(12), 2704–2711 (2017) (in Chinese)Google Scholar