A two-stage multi-AGV task allocation method based on capacity constrained K-means clustering and improved genetic algorithm for smart logistics parks

Abstract

Aiming at the problems of low task allocation efficiency, unbalanced vehicle load and high total driving cost faced by multi-AGV system in intelligent logistics park under multi-task collaborative operation scenario, a multi-AGV task allocation method combining capacity constraint clustering and improved genetic algorithm is proposed. Firstly, aiming at the problem that the traditional K-means only considers the spatial distance and is difficult to meet the vehicle load constraints, a capacity-constrained K-means pre-clustering method is constructed to realize the spatial compact division of task points under the condition of meeting the vehicle capacity constraints. Then, the optimization of task access sequence in each cluster is modeled as a traveling salesman problem, and an improved genetic algorithm for permutation coding is designed. The algorithm improves the feasibility, convergence speed and stability of the solution through high-quality hybrid initialization, crossover strategy adapted to permutation coding, composite mutation mechanism, hierarchical local search and elite retention mechanism. Simulation experiments show that the proposed method is superior to the benchmark method in terms of task load balancing, clustering feasibility and solving performance on the TSPLIB standard test set, and can effectively reduce the total driving cost of the system while taking into account the capacity constraints. The research results can provide effective technical support for task allocation in multi-AGV scheduling of smart logistics parks.