A landscape-aware multi-objective ant colony optimization framework for urban waste collection vehicle routing

Abstract

Municipal solid waste collection routing has traditionally been optimized for transportation efficiency while neglecting the heterogeneous visual, acoustic, and ecological disturbances imposed on landscape sensitive urban regions. To address this limitation, a unified landscape-aware multi-objective optimization framework is proposed in which continuous environmental perception is integrated directly into routing decision-making. A multi-criteria landscape sensitivity field $$\:\left(S\right(x,y\left)\right)\:$$ is constructed using Gaussian radial basis potentials to represent heterogeneous urban sensitivities, and a dimensionless Landscape Conflict Index (LCI) is introduced to quantify the environmental intrusion intensity of routing solutions. Based on this formulation, a Landscape-aware Ant Colony Optimization (LA-ACO) algorithm is developed by embedding generalized landscape-aware cost into both heuristic transition and pheromone reinforcement mechanisms, with iterative 2-opt refinement incorporated to improve convergence stability. Extensive experiments involving convergence analysis, robustness evaluation, cross instance validation, and a Zhengzhou inspired real-world case study demonstrate that LA-ACO consistently outperforms conventional metaheuristic baselines in balancing operational efficiency and environmental preservation. Compared with Standard ACO, the proposed framework achieves a 9.4% reduction in generalized routing cost (significant in nine of ten paired seeds; one-sided paired Wilcoxon p = 0.002) and produces a Pareto hypervolume 90.5% larger than that of NSGA-II (198.4 vs. 104.1). At the empirically identified knee point operating regime ( $$\:{\alpha\:}^{*}=2$$ ), landscape conflict is reduced by 31.1% with a 13.9% increase in route distance, corresponding to a marginal benefit of 2.2% points of LCI reduction per percentage point of distance cost. The proposed framework establishes a generalized spatially aware optimization paradigm for sustainable urban logistics and provides quantitatively interpretable decision support for environmentally coordinated smart city governance.