Cooperative dispatching of port integrated energy and ferry logistics systems under interrelated cross-domain uncertainties

To address potential threats posed by cross-domain uncertainties in integrated energy and ferry logistics to port operations, this paper proposes a cooperative dispatching strategy. Uncertainties in integrated energy production and ferry logistics operations, along with their interrelations, are first modeled. A fluctuation range prediction (FRP) method, utilizing convolutional neural networks and bi-long short-term memory networks (CNN-Bi-LSTM), is developed to reduce conservatism and improve the accuracy of uncertainty modeling. The interval formulation (IF) method is applied to reduce variable dimensionality and enhance solvability. The dispatching of port integrated energy and ferry logistics systems under these uncertainties is formulated as two-stage robust distributed optimization problems. A cooperative optimization approach is introduced by integrating the column-and-constraint generation algorithm with the adaptive alternating direction method of multipliers (C&CG-AADMM) to achieve decentralized cooperation and robust optimal dispatches. This approach improves system robustness and cooperation efficiency by supporting independent robust yet interactive optimizations while preserving the decision-making autonomy of each system. Simulation results demonstrate the effectiveness of the proposed strategy in co-enhancing energy and logistics operational benefits amidst cross-domain and interrelated uncertainties.