A Many-Objective Optimization Algorithm for Multi-Vehicle Candidate Path Combinations

Jiang, Hongyan

doi:10.1142/s0218194026500245

Return

A Many-Objective Optimization Algorithm for Multi-Vehicle Candidate Path Combinations

2026-04-01

PRE

DOI：10.1142/s0218194026500245

Original

Original request for help

Save

Abstract

En 中文

As the scale of multi-vehicle parallel operations in smart logistics systems expands and warehouse environments become increasingly complex, involving the coordination between multiple vehicles and the interaction of spatiotemporal constraints, these complex environmental factors make path combination optimization particularly challenging. With the increase in the number of objectives, the optimization problem transforms from traditional multi-objective optimization problems (MOPs) to many-objective optimization problems (MaOPs), where the solution space expands exponentially, leading to the occurrence of the dimension catastrophe. To address these challenges of many-objective optimization and complex spatiotemporal constraints, this paper constructs a multi-vehicle collaborative path combination optimization model that comprehensively balances system total distance, potential conflict risk, overall operational cost and task duration. The specific research approach is as follows: First, an improved K Shortest Paths (KSP) technique based on a composite heatmap integrating candidate path popularity and vehicle diversity penalties is developed to generate pre-selected collaborative scenarios that balance distribution diversity and safety. Second, a many-objective path combination model is constructed to balance total distance, potential conflict rate, total cost and task duration. Finally, an efficient discrete many-objective particle swarm optimization algorithm is designed, incorporating a two-archive coordination mechanism with attractor and repeller forces. Experiments demonstrate that the proposed algorithm delivers well-balanced scheduling solutions with controllable conflict risks for complex many-objective path combination models, achieving favorable convergence and diversity performance results.

Keywords:

K Shortest Paths

route planning

path diversity

many-objective optimization

interaction force