⚡ electrical engineering

Multi-Objective Ship Route Planning in Complex Ocean Environments Using an Environment-Aware Adaptive Ant Colony Optimization Algorithm

This study proposes an Environment-Aware Adaptive Multi-Objective Ant Colony Optimization (EA-MOACO) algorithm that integrates coupled wind–wave–current conditions and dynamic parameter adjustments to outperform existing methods in minimizing both route length and energy consumption for ship navigation in complex marine environments.

Original authors: Qiang Yuan, Yujie Jiang, Junsong Lei, Yu Yang, Baokui Wang, Jun Ye, Fei Dong

Published 2026-08-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Qiang Yuan, Yujie Jiang, Junsong Lei, Yu Yang, Baokui Wang, Jun Ye, Fei Dong

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Navigating the open ocean is a constant negotiation with a restless, invisible world. Unlike driving on a highway, where the road is fixed and the weather is merely a backdrop, a ship moving through the sea is physically pushed and pulled by the wind, waves, and currents that surround it. These forces do not just make a journey uncomfortable; they fundamentally alter how much fuel a vessel needs to reach its destination. A route that looks shortest on a map might actually be the most expensive if it fights against a strong headwind or a powerful current, while a slightly longer path that rides with the flow could save tons of fuel and reduce dangerous emissions. For decades, computer systems helping captains plan these journeys have often treated the ocean as a static map, focusing only on distance while ignoring the dynamic energy costs of the environment. As ships become more automated and the global push for greener shipping grows, the need to account for these invisible forces has become critical.

Researchers at Zhejiang Ocean University and their colleagues have developed a new computer method designed to solve this specific problem. They created a system that does not just look at the distance between two points, but actively "feels" the ocean conditions along the way. The team built a digital model of the sea using real historical weather data, simulating how wind, waves, and currents interact with a ship's hull. They then combined this realistic ocean model with a smart search algorithm inspired by how ants find food. In nature, ants leave chemical trails to guide others to a food source; the researchers adapted this behavior so that their digital "ants" leave trails not just for the shortest path, but for the most energy-efficient path. The system constantly adjusts its search strategy based on how difficult the environment is, allowing it to explore widely when conditions are chaotic and focus tightly when the way is clear.

The core of this new approach is a deep integration of environmental awareness. Instead of assuming the water is calm, the system pulls in data on wind speed and direction, wave height, and current strength. It calculates exactly how much extra resistance a ship would face in these conditions. For instance, it knows that a ship moving against a strong wind needs more power, and that a ship moving with a current can coast more easily. By feeding this detailed physics into the planning algorithm, the system can identify routes that might look longer on a map but require significantly less fuel to travel. The researchers tested this method on a simulated journey across a section of the ocean off the coast of China, using a model of a medium-sized cargo ship. They compared their new system against several other existing planning methods, including standard ant colony algorithms and other advanced optimization techniques.

The results of these simulations showed a clear advantage for the new method. It consistently found routes that were both shorter and less energy-intensive than those found by the other systems. In the tests, the new algorithm reduced the average travel distance by more than two percent and cut the energy consumption by a similar margin. While these percentages might seem small, in the context of global shipping where vessels travel thousands of miles, this translates into massive savings in fuel and a significant reduction in greenhouse gas emissions. The system also proved to be very stable; when the researchers ran the simulation fifty times with different random starting points, the results remained consistent, showing that the method does not rely on luck to find a good route. It successfully balanced the competing goals of speed and fuel economy, offering a set of viable options rather than forcing a single choice.

What makes this work particularly robust is its ability to adapt to changing conditions. The researchers found that as the simulated ocean became more violent, with stronger winds and higher waves, the system did not panic or fail. Instead, it adjusted its internal settings to prioritize finding a safe, energy-efficient path over simply finding the shortest one. It recognized that in rough seas, the cost of fighting the elements is too high, so it would naturally choose a route that might be slightly longer but much calmer. This adaptability suggests that the system could handle real-world unpredictability, where weather conditions can shift rapidly. The study explicitly ruled out the idea that simple distance-based planning is sufficient for modern shipping, demonstrating that ignoring environmental factors leads to suboptimal and wasteful routes.

The researchers validated their findings through rigorous testing, running the simulations on a standard computer workstation. They did not just look at the final route but analyzed how the algorithm performed at every step, ensuring it was not getting stuck in local solutions or missing better options. The new method outperformed other sophisticated algorithms, including those that use deep learning or hybrid techniques, by finding a better balance between exploring new possibilities and refining the best known paths. The study concludes that by treating the ocean as a dynamic, energy-consuming partner rather than a passive background, ship route planning can become significantly more efficient. This approach offers a practical, computationally efficient way to guide ships through complex waters, promising a future where maritime transport is safer, cheaper, and kinder to the planet. The work stands as a simulation-based proof of concept, showing that when computers are taught to respect the physics of the sea, they can chart a better course for us all.

Technical Summary: Multi-Objective Ship Route Planning in Complex Ocean Environments Using an Environment-Aware Adaptive Ant Colony Optimization Algorithm

1. Problem Statement

Ship route planning in complex marine environments faces significant challenges due to the dynamic and uncertain nature of wind, waves, and ocean currents. Traditional path planning methods, such as Dijkstra, A*, and standard Ant Colony Optimization (ACO), often rely on geometric distance as the primary metric. These approaches frequently neglect the impact of environmental disturbances on vessel energy consumption and maneuverability. Furthermore, existing algorithms often suffer from:

Lack of Environmental Awareness: Inability to dynamically perceive and adapt to real-time wind-wave-current interactions.
Single-Objective Limitations: Difficulty in simultaneously optimizing conflicting objectives, such as minimizing route length and reducing energy consumption.
Premature Convergence: Fixed parameters in traditional ACO often lead to suboptimal solutions in complex search spaces, failing to balance global exploration and local exploitation effectively.

2. Methodology

The authors propose the Environment-Aware Adaptive Multi-Objective Ant Colony Optimization (EA-MOACO) algorithm. This framework integrates realistic environmental modeling with an adaptive search strategy to optimize ship routes under coupled wind-wave-current conditions.

2.1 Environmental and Energy Modeling

Coupled Wind-Wave-Current Model: The study utilizes ERA5 reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) to construct a realistic environmental field.
- Wind: Derived directly from ERA5 u10 and v10 components.
- Waves: Significant Wave Height (SWH) and wave direction are modeled based on empirical formulations.
- Currents: Constructed using a wind-drift approximation model (surface current velocity $\approx$ 2% of local wind speed) due to the lack of direct current data in the selected ERA5 subset.
Grid Representation: The navigation area is discretized into a 2D grid with an 8-neighbor connectivity structure. A collision-constrained transition strategy is employed to prevent corner collisions without artificially expanding obstacles.
Ship Energy Consumption Model: A comprehensive model calculates total resistance ( $R$ $R$ ) as the sum of hydrostatic friction resistance ( $R_f$ $R_{f}$ ), wind resistance ( $R_{wind}$ $R_{w in d}$ ), and wave-induced resistance ( $R_{wave}$ $R_{w a v e}$ ).
- Speed Correction: Effective ship speed ( $V_{eff}$ ) is adjusted based on the angle between the ship's heading and ocean currents.
- Energy Calculation: Energy consumption is derived from total resistance, effective speed, and voyage duration, accounting for the specific vessel's hydrodynamic and aerodynamic properties.

2.2 The EA-MOACO Algorithm

The core of the proposed method involves three adaptive mechanisms:

Environmental Complexity Assessment:
- A composite index ( $C$ ) is calculated based on obstacle density and the normalized distance between start and end points.
- This index categorizes the environment into Simple, Moderate, or Complex, guiding the initial parameter settings.
Adaptive Parameter Control:
- Initialization: The pheromone importance factor ( $\alpha$ ) and heuristic importance factor ( $\beta$ ) are initialized based on environmental complexity. Complex environments favor higher $\beta$ (heuristic/exploration), while simple environments favor higher $\alpha$ (pheromone/exploitation).
- Dynamic Adjustment: During iterations, $\alpha$ and $\beta$ are dynamically updated. The strategy shifts from exploration (high $\beta$ ) in early stages to exploitation (high $\alpha$ ) in later stages, balancing global search and local convergence.
Energy-Aware Multi-Objective Search:
- Improved Heuristic Function: The heuristic value ( $\eta_{ij}$ ) incorporates both geometric distance and normalized energy consumption, weighted by a factor ( $\lambda$ ) that scales with environmental complexity.
- Dynamic Weighting: A hierarchical framework adjusts the weights of path length and energy consumption based on iteration progress and environmental complexity. Early iterations prioritize finding feasible paths (length), while later iterations refine for energy efficiency.
- Pareto Optimization: The algorithm employs Pareto non-dominated sorting to maintain a diverse set of trade-off solutions.
- Ideal-Point Decision Making: The final route is selected from the Pareto front by minimizing the Euclidean distance to an ideal point (0,0) in the normalized objective space.

3. Key Contributions

The paper identifies four primary contributions:

Coupled Modeling: Establishment of a wind-wave-current coupled ocean environment model using ERA5 data, integrated with a ship energy-consumption model that accounts for current-induced speed correction and environmental resistance.
Adaptive Mechanism: Development of an environmental-complexity-driven adaptive parameter adjustment mechanism that dynamically regulates pheromone and heuristic factors to suit varying search difficulties.
Multi-Objective Framework: Proposal of a dynamic multi-objective optimization framework that integrates route length and energy consumption using adaptive weighting, Pareto dominance, and ideal-point decision-making.
Comprehensive Validation: Extensive experimental validation including ablation studies, robustness analyses, environmental-intensity tests, and comparative evaluations against state-of-the-art algorithms.

4. Experimental Results

The EA-MOACO algorithm was evaluated against ACO, PSO, MOPSO, NSGA-II, PSO-ACO, and DL-DACO using a medium-sized cargo vessel model in a simulated environment (121°E–123°E, 29°N–31°N).

Performance Metrics: EA-MOACO outperformed all comparison algorithms in terms of Hypervolume (HV), Inverted Generational Distance (IGD), path length, and energy consumption.
- Path Length: In ablation studies, the complete EA-MOACO framework reduced the average path length from 36.104 to 35.340, corresponding to an improvement of 2.12% compared to the baseline ACO. In comparative experiments against standard ACO (Table 9), EA-MOACO achieved a path length of 35.524 versus 36.036 for ACO.
- Energy Consumption: In ablation studies, the framework reduced average energy consumption from 9.072×10⁶ to 8.881×10⁶, corresponding to a reduction of 2.11% compared to the baseline. In comparative experiments against standard ACO (Table 9), EA-MOACO achieved 8.876×10⁶ versus 9.037×10⁶ for ACO.
- Multi-Objective Quality: In ablation studies, the HV of the complete algorithm improved by 6.58% compared to the baseline experiment, and the IGD value decreased by 4.70% compared to the baseline ACO. In comparative experiments, EA-MOACO achieved the highest HV and lowest IGD among all tested algorithms, demonstrating superior diversity and convergence.
Ablation Studies: Stepwise addition of modules (environmental modeling, parameter adaptation, multi-objective planning) demonstrated progressive improvements in all metrics, validating the necessity of each component.
Robustness: Independent repetition experiments (50 runs) showed low coefficients of variation for path length (2.80%) and energy consumption (2.72%), confirming the algorithm's stability.
Environmental Intensity: Tests under varying environmental intensities (weak to extreme) showed that while energy consumption increased with environmental severity, the planned path length remained stable, demonstrating the algorithm's ability to adapt to harsh conditions without excessive detours.

5. Significance and Claims

The paper claims that the EA-MOACO algorithm provides an efficient and practical framework for intelligent ship route planning in complex marine environments. Its significance lies in:

Deep Integration: Successfully integrating real-world environmental data (ERA5) directly into the optimization logic, moving beyond geometric-only planning.
Adaptability: The ability to dynamically adjust search strategies based on environmental complexity and iteration progress, mitigating the risk of premature convergence common in traditional ACO.
Balanced Optimization: Effectively resolving the trade-off between navigation efficiency (shortest path) and economic efficiency (lowest energy consumption), offering decision-makers a diverse set of Pareto-optimal routes.
Engineering Applicability: The algorithm achieves high-quality results without the computational overhead of deep learning training (unlike DL-DACO), making it suitable for practical engineering applications in autonomous shipping.

The authors conclude that this approach addresses the critical limitations of current methods in handling coupled environmental factors and multi-objective constraints, offering a robust solution for green and intelligent shipping. Future work is suggested to include real-time dynamic updates, integration of maritime traffic regulations, and the exploration of deep reinforcement learning and digital twin technologies.