Using the unbiased grey Markov chain optimization model to forecast passenger flows through existing urban rail transit stations

To accurately forecast the passenger flows of existing stations in urban rail transit and provide a scientific basis for subway operation management departments, a passenger flow prediction model optimized by an unbiased grey Markov chain was constructed. Using 18 months of average daily passenger flow data from Shilaoren Bathing Beach Station on Qingdao Metro Line 2, a grey prediction model was established and tested. The data were classified into six states according to their relative errors for Markov state modeling, and the classification results were validated. A state transition probability matrix and the corresponding k-step transition probabilities were then established. The unbiased grey Markov chain prediction model was used to predict the passenger flow data of existing stations in the next 3 months. By comparing the prediction results with the actual passenger flow data, it was found that the prediction error after optimization by the unbiased grey Markov model was controlled within 10%, which corresponded well with the actual characteristics of urban rail transit passenger flows. This model is suitable for predicting the passenger flows of existing urban rail transit stations and can provide an effective reference for managing subway operations and ensuring safety.