Design of low-carbon planning model for vehicle path based on adaptive multi-strategy ant colony optimization...

In contemporary transportation systems, the imperatives of route planning and optimization have become increasingly critical due to vehicles' burgeoning number and complexity. This includes various vehicle types, such as electric and autonomous vehicles, each with specific needs. Additionally, varying speeds and operational requirements further complicate the process, demanding more sophisticated planning solutions. These systems frequently confront myriad challenges, including traffic congestion, intricate routes, and substantial energy consumption, which collectively undermine transportation efficiency, escalate energy usage, and contribute to environmental pollution. Hence, strategically planning and optimizing routes within complex traffic milieus are paramount to enhancing transportation efficacy and achieving low-carbon and environmentally sustainable objectives. This article proposes a vehicle path low-carbon planning model, Adaptive Cooperative Graph Neural Network (ACGNN), predicated on an adaptive multi-strategy ant colony optimization algorithm, addressing the vehicle path low-carbon planning conundrum. The proposed framework initially employs graph data from road networks and historical trajectories as model inputs, generating high-quality graph data through subgraph screening. Subsequently, a graph neural network (GNN) is utilized to optimize nodes and edges computationally. At the same time, the global search capability of the model is augmented via an ant colony optimization algorithm to ascertain the final optimized path. Experimental results demonstrate that ACGNN yields significant path planning outcomes on both public and custom-built datasets, surpassing the traditional Dijkstra's shortest path algorithm, random graph network (RGN), and conventional GNN methodologies. Moreover, comparative analyses of various optimization methods on the custom-built dataset reveal that the ant colony optimization algorithm markedly outperforms the simulated annealing algorithm (SA) and particle swarm optimization algorithm (PSO). The method offers an innovative technical approach to vehicle path planning and is instrumental in advancing low-carbon and environmentally sustainable goals while enhancing transportation efficiency.