Smart Waste Collection Via PID Control and Hybrid Metaheuristic Optimization

Abstract

This paper presents a simulation-only baseline for municipal waste collection that integrates discrete-time PID bin-level control with metaheuristic logistics in MATLAB/Simulink R2025a. The truck’s “approach + service” dynamics are discretized with a zero-order hold at a sampling time of 0.001 s over a 200 s horizon. A two-degree-of-freedom parallel discrete PID controller is applied with tuned parameters, including proportional, integral, and derivative gains, along with a derivative filter. The urban environment is represented as a [0,100]×[0,100] m Euclidean plane containing 10 bins with randomly generated coordinates. At each decision epoch (every 10 s), a temporary depot is selected using Simulated Annealing (initial temperature 100, final temperature 1e−3, cooling rate 0.9, 1000 iterations), and a closed tour is computed by Ant Colony Optimization (20 ants, 100 iterations, α=1, β=5, ρ=0.5, Q=100). Bin inflow is evaluated under constant, ramp, and sinusoidal profiles, with eligibility triggered at a fill threshold of 0.8. Travel times are based on a vehicle speed of 40 km/h and a per-bin service time of 10 s. For a representative constant-load trial, the selected depot was located at (71.6734, 67.1613) m, and the best tour length was 356.0113 m. The framework reports key performance indicators including overflow rate, service latency, and travel distance, and is designed to enable comparative evaluation against greedy or untuned baselines without relying on hardware or field data.