Multi-agent systems for optimising smart energy clusters: A case study on cost and emission reduction in industrial seaport facilities

This study investigates the use of multi-agent systems (MAS) to optimise energy flow in industrial seaport clusters, with an emphasis on reducing operational costs, minimising carbon emissions, and increasing renewable energy utilisation. The proposed framework combines decentralised control via MAS, dynamic load scheduling, and smart grid interaction across three operational strategies: individual building-level optimisation, cluster-level coordination, and peer-to-peer energy trading. Simulation results from a UK-based pilot site indicate that cluster-level optimisation achieves an average 60% cost saving and 30% emission reduction. The peer-to-peer strategy enables up to 90% renewable self-consumption, reducing grid dependence by 30–35%. In contrast, the individual strategy remains sensitive to demand fluctuations, with grid dependency rising to 70% under high-load conditions due to limited energy sharing and battery saturation. These findings demonstrate the scalability and adaptability of MAS-based energy frameworks in industrial contexts. By critically evaluating performance under variable operational scenarios, this study offers practical insights for sustainable energy management and industrial decarbonisation pathways. • Multi-agent model optimises energy flow, costs, and emissions in industrial seaports. • Cluster optimisation outperforms individual methods, saving 60% of energy costs. • Peer-2-peer energy boosts renewable use to 90%, reducing grid reliance to 30%–35%. • Multi-agent framework tested on real information, scalable to industrial sectors.