Study on Solving Dynamic Vehicle Routing Problem Based on Deep Learning for Cost Control

View at publisher Summarise with VIRAT
Authors: Jingbo Yu
Cited by: 0
References: 12

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Aiming at the cost control requirements in the Dynamic Vehicle Routing Problem (DVRP) and the problems of poor target orientation, slow convergence speed, and many path inflection points existing in the traditional Rapidly-exploring Random Tree (RRT) algorithm for path planning, this paper proposes an improved RRT algorithm integrated with deep learning concepts to achieve efficient path planning and cost optimization for vehicles in dynamic scenarios. The improved algorithm combines asymptotic region sampling with a target-biased sampling mechanism, which effectively avoids the reverse search of random sampling points, improves the utilization rate of the sampling space, and strengthens the directionality of sampling, laying a foundation for the real-time performance and accuracy of dynamic path planning. It introduces the dynamic circle regularization method to ensure that vehicles can effectively avoid collisions and pass through narrow areas smoothly in dynamic environments. Additionally, it adopts a path redundant point pruning method based on the triangle inequality, combined with the cubic B-spline smoothing technology to optimize the path shape, thereby reducing the energy consumption and time cost of path travel. Finally, through MATLAB simulation experiments and comparative analysis, it is verified that the proposed improved algorithm is not only significantly superior to the traditional RRT algorithm in terms of the number of searches and planning time, but also the planned path is more in line with the actual needs of the dynamic vehicle routing problem. It can effectively reduce the comprehensive cost during vehicle travel and provide a feasible solution for cost control in the dynamic vehicle routing problem.

Fields of Study & Subjects

Robotic Path Planning Algorithms Vehicle Routing Optimization Methods Reinforcement Learning in Robotics

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
477–482
Publication date
28 January 2026
Language
EN