This example chooses a route through a small delivery graph. Each route segment has a normal cost and a risk value. The question is not simply “what is shortest?” but “what is best after risk is priced in?”
The program gives every edge a risk-adjusted score: delivery cost plus a risk penalty. The risk penalty is the edge risk multiplied by the configured risk weight.
It then uses Dijkstra’s shortest-path algorithm. Dijkstra expands the currently cheapest known partial route first. Once the destination is removed from the queue, the selected route is optimal for the weighted graph.
The selected route goes through DepotB and LabD. A route through DepotC looks cheaper early, but it carries enough risk that it loses after the risk penalty is applied. The output separates raw cost, risk sum, and risk-adjusted score so the trade-off is visible.
The trust gate verifies that edge costs and risks are non-negative, at least one path reaches the goal, the selected path starts and ends correctly, the path appears in the enumerated simple paths, and no enumerated simple path has a lower risk-adjusted score.
From the repository root:
node examples/dijkstra_risk_path.js