Blockchain-enhanced reinforcement learning: A sustainable solution for optimized municipal solid waste management

Municipal Solid Waste (MSW) management remains a critical urban challenge, with existing systems often constrained by inefficiencies, high operational costs, and environmental impacts. While prior studies have applied optimization or digital technologies separately, there is still a lack of integrated approaches that combine operational efficiency with transparency and sustainability. This study addresses this gap by proposing a novel framework that integrates Reinforcement Learning (RL) for route optimization with blockchain for secure and transparent transaction recording in MSW management. The methodology employs a synthetic but realistic dataset of 1000 collection points, incorporating waste type, volume, distance, fuel use, and operational costs. The RL model processes these as input states to determine optimal routing actions that minimize Travel Distance (TD) and emissions while improving efficiency. The resulting outputs optimized routes, waste volumes, costs, and emissions are recorded on a blockchain, ensuring transparency and immutability. The experimental design compares the RL-blockchain framework against traditional collection methods, with performance evaluated through Material Savings (MS), energy savings, CO₂ Emission Reductions (CER), and cost efficiency. Results show that the integrated system achieves MS of 395,342.6 kg, energy savings of 142,763.8 kWh, carbon emission reductions of 35,589.4 kg CO₂eq, and Cost Savings (CS) of USD 2,812,476.3 relative to traditional methods. Overall, this study contributes a scalable, transparent, and sustainable decision-support model for urban waste management, offering practical applications for municipalities seeking to reduce costs and environmental burdens while enhancing accountability.