One lake. Many possible missions.
Each point represents one feasible nondominated mission found in the final population. Select a point to see how the trade-off between mission time, longest individual route and Cyanobacteria-density coverage changes the planned ASV trajectories.
Open full-page explorer ↗Loading recorded missions…
Pareto front
Hover or select a point. Use arrow keys when the plot is focused.
Planned trajectories
Plan view · equal x/y scale · coordinates in metres
- Mission duration
- Longest individual path (3D)
- Cyanobacteria-density coverage
Trade-off descriptions are relative to this front. “Balanced” gives equal weight to all three normalized objectives; it is not universally best.
About the data & how to read it
Only feasible, nondominated final solutions are shown, sampled from the median final-hypervolume repeat of each experiment. Hypervolume describes an entire front, never an individual mission. Two-ASV runs finish at generation 200; three-ASV runs at 500. These are different optimisation budgets, so this view is not a controlled fleet-performance comparison.
Mission duration is the last boat’s arrival. Longest path is the greatest individual 3D spline arc length, while the map shows horizontal projections. Displayed Cyanobacteria-density coverage is exactly the negative of the saved third minimization objective: coverage = −objs(3). It sums each particle’s greatest Gaussian contribution across all boat paths, avoiding double-counting overlaps. It is a model score, not measured concentration.
The background is a horizontal Gaussian projection of recorded particles at mission start, not cumulative Cyanobacteria-density coverage or an animated bloom. The STL outlines are the complete saved simulation geometries, including shoreline holes; no geographic imagery is added. Different groups are displayed separately, with axes rescaled to each front.