Author(s)

Le Yang, Mengtian Li, Tianle Shen

Corresponding Author

Le Yang

Abstract

This research focuses on the optimization of campus bike-sharing scheduling and parking point layout by developing a spatiotemporal collaborative optimization model. Through extreme value truncation method for determining bicycle distribution, mixed integer programming for scheduling models, entropy weight-K-means algorithm for parking point layout optimization, and genetic algorithm for dynamic path planning of faulty bicycles, we achieved significant improvements. Results show that the scheduling model successfully reallocated 1,240 bicycles across 7 time periods, optimized parking points coverage increased from 68% to 92% with efficiency scores rising to 0.75, and the adaptive genetic algorithm effectively balanced transportation efficiency and risk control. Model robustness was confirmed through sensitivity analysis and spatial autocorrelation verification, demonstrating potential for extension to urban micro-transportation systems.

Keywords

Bike-Sharing Scheduling, Mixed Integer Programming, Entropy Weight-AHP Model, K-Means Clustering, Genetic Algorithm, Spatiotemporal Window Matching