Objective To address the substantial load challenges of marine air conditioning systems, this paper proposes integrating ice storage cooling technology into the air conditioning system. A Multi-strategy Improved Dung Beetle Optimization algorithm (MIDBO) is employed to optimize the operational strategy.

Method First, the Hammersley sequence is adopted to initialize the population, enhancing uniform coverage of the search space. Next, the Lévy flight mechanism is introduced to enhance the algorithm's jump behavior and global search ability. Finally, an adaptive spiral search strategy is integrated to balance global exploration with local exploitation. Through this multi-strategy fusion, MIDBO effectively improves the operational efficiency and economic performance of marine ice storage cooling systems.

Results Simulation experiments show that MIDBO outperforms classic optimization algorithms such as GWO, PSO, and SSA across all 10 benchmark test functions, exhibiting stronger optimization capability and convergence accuracy. Applied to the marine ice storage cooling system, the MIDBO strategy reduces daily fuel consumption to 3338.6 kg, saving 36.6 kg compared to the baseline scheme. The required ice storage tank volume is only 10.77 m³, with an investment cost of 41449.43 USD, annual cost savings of 7171.30 USD, and a payback period of 5.20 years, significantly better than the traditional peak shaving strategy's 5.82 years and results from other optimization algorithms. Environmentally, the MIDBO scheme achieves an annual CO₂ emission reduction of 21.12 tons, a reduction ratio of 1.08%, surpassing that of other algorithms. Sensitivity analysis reveals that fuel price is the dominant factor affecting system economics, having a far greater impact than interest rate or charging temperature. Unlike traditional "peak shaving and valley filling" strategies, MIDBO achieves optimal system resource allocation by precisely controlling the load to keep diesel generators operating within their high-efficiency range whenever possible.

Conclusion The proposed MIDBO algorithm significantly optimizes the operational strategy of marine ice storage cooling systems, enhancing energy utilization efficiency, reducing operating costs, and minimizing environmental impact.