Optimization of Train Speed Profiles for a Metro Transit System by Genetic Algorithms

Traditional Automatic Train Operation (ATO) algorithms are generally designed based on single-train operation with the objective of improving the speed profile of a single train to reduce mechanical energy consumed under operational constraints. For many electrified rail transit systems where energy cost is calculated at the substation level, minimizing energy consumption needs to consider the reuse of regenerative energy from neighboring trains in the same power section. If regenerative energy is considered, when two trains are moving in the same power section, one train adjusts its speed to a different speed profile according to the position, speed and regeneration potential of the other train to reuse the maximum amount of regenerated energy. With the dual objectives of maintaining schedule requirements and optimizing energy efficiency, this paper analyses dynamic and electric performance of two opposing trains operated in the same DC power section. Genetic algorithms have been applied to search for the optimal train speed profiles. Tractive/braking efforts of both trains and energy cost at substation level are defined as strings of chromosome and the fitness function respectively. Simulation through Visual C++ platform demonstrates that the algorithm can provide optimal train speed profiles with better energy performance while satisfying operational constraints. Different synchronization times have different optimization ratios. This research will help facilitate development of on-board train control system logic to analyze energy flow in a multi-train network and reduce overall energy consumption.