Every grid cell holds a movement resistance r (1 = open habitat, up to 50 = a near-impassable barrier such as a road or built-up patch), generated from layered noise plus scattered barrier blobs. Two connectivity models run on the same resistance surface, both used in real conservation GIS tools (Circuitscape, Linkage Mapper):
Cost distance: D(p) = min over paths Σ (r_i+r_j)/2 · step
Least-cost corridor: keep cells where
D_source(p) + D_dest(p) ≤ D_min · (1 + tolerance)
The least-cost path is the single cheapest route (Dijkstra's algorithm on an 8-connected grid, edge weight = average resistance of the two cells it crosses). Widening the tolerance slider reveals every near-optimal alternative route as a corridor band, not just one line — exactly how landscape planners size a conservation corridor.
Circuit theory (Kirchhoff's current law):
conductance g_ij = 1 / avg(r_i, r_j)
voltage V_i = Σ g_ij·V_j / Σ g_ij (relaxation)
current I_ij = g_ij·(V_i − V_j)
R_eff = V / I_total (V=1 fixed at source, 0 at destination)
Fixing the source patch at 1 volt and the destination patch at 0 volts and solving Laplace's equation by relaxation gives a diffuse current-flow map: current spreads across every viable route in proportion to how conductive it is, revealing braided or redundant corridors that a single least-cost line hides. The effective resistance between the two patches is a real, physically meaningful measure of how well-connected they are — lower is better connected.
- Barrier density + Regenerate — builds a new resistance surface and re-solves both models.
- Corridor width tolerance — widens or narrows the least-cost corridor band around the optimal path.
- Disperser speed — animates tokens moving along the least-cost path.