Multi-Objective Gannet Optimization Algorithm for Dynamic Passenger Flow Allocation in Train Operation Plan Optimization

This paper proposes a multi-objective Gannet Optimization Algorithm (MOGOA) to address the issue of unbalanced train occupancy rates in railway train operation planning. MOGOA employs an adaptive multi-population co-evolutionary strategy to balance exploration and exploitation, utilizing a non-dominated sorting algorithm based on crowding distance to select parent and child samples. These samples serve as initial solutions for subsequent iterations. A novel maximin fitness function guides the iterative update of the global optimal position. MOGOA is applied to the train operation planning problem with dynamic passenger flow allocation feedback. It collaboratively optimizes the number of train operations, sections, and stops to reduce costs, balance occupancy rates, minimize travel time, and enhance travel satisfaction. The practical applicability of MOGOA in optimizing train operation plans based on dynamic passenger flow allocation is significant.