Bi-level convex optimization speed planning algorithm for pure electric buses under multi-factor urban road conditions

Energy saving is crucial in eco-driving research of pure electric buses, and its effectiveness is highly dependent on the speed planning. The current speed planning methods use dynamic programming to find the optimal solution, but they become less suitable when multi-factor conditions introduce comp...

Full description

Saved in:
Bibliographic Details
Published in:International journal of green energy Vol. 22; no. 13; pp. 2813 - 2825
Main Authors: Li, Chenyang, Lin, Qiang, Chen, Longxiang, Zhu, Qingyuan, Shao, Guifang
Format: Journal Article
Language:English
Published: Taylor & Francis 03.10.2025
Subjects:
convex optimization algorithm
Energy saving
multi-factor conditions
pure electric buses
speed planning
ISSN:1543-5075, 1543-5083
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Energy saving is crucial in eco-driving research of pure electric buses, and its effectiveness is highly dependent on the speed planning. The current speed planning methods use dynamic programming to find the optimal solution, but they become less suitable when multi-factor conditions introduce complex constraints, leading to more computation complexity and time. To reduce the computational complexity, we proposed a bi-level convex optimization speed planning algorithm (Bi-COA) for pure electric buses under multi-factor urban road conditions. This method includes the constraint layer and solution layer. In the constraint layer, we constructed a nonlinear constraint model that considers multiple factors such as load variation, road gradient, and traffic signals, and transformed them into linear constraints through convexification. We then employ the average speed to transform the energy-optimal cost function into a quadratic function, which can reduce the computational complexity. In the solution layer, the Multi-Objective Evolutionary Strategy Kernel (MOESK) is utilized to obtain the optimal driving speed of the pure electric bus under multiple constraints. The results of the experiments indicate that the proposed method can save 71.99% of energy consumption compared with human drivers. Compared with the dynamic programming method, its average solution efficiency is improved by 40%, which greatly reduces the calculation time.
ISSN:1543-5075
1543-5083
DOI:10.1080/15435075.2025.2471999