AI ORIGINS · LESSON 08

Rescue Grid: The A* Mission

Guide the rover to the beacon using A* while terrain, fuel pressure, and mission rules change in real time.

Paint Tool

Build your grid, then run the mission.

Rover (Start)
Beacon (Goal)
Debris (Blocked)
Sand (Cost 3)
Open Set
Closed Set
Final Path

Instrument Panel

Fuel Spent g(n)0.00
Estimated Distance h(n)0.00
Total Mission Cost f(n)0.00

Mission Metrics

Nodes Expanded0
Final Path Cost--
Time to Solve--
Optimality CheckPending

Heuristic Comparison

Same grid, three heuristics.

Heuristic	Nodes	Cost	Time
Manhattan	--	--	--
Euclidean	--	--	--
Zero (Dijkstra)	--	--	--

Mission Brief

You are testing how A* makes decisions under different constraints and mission rules.

Narrative Context

Navigate across the Martian surface to reach the rescue beacon while avoiding craters and slow sand zones.

What To Experiment With

HeuristicsCompare speed and expansions across Manhattan, Euclidean, and Dijkstra.
Mode RulesTest Fuel Critical and see when a valid route becomes impossible.
Step ReasoningUse Step Mode to inspect how open and closed sets evolve node by node.

Reflection

Prompt 1Which heuristic solved your map with the fewest node expansions, and why?
Prompt 2What changed when you switched to Fuel Critical mode?
Prompt 3In Dynamic Obstacles mode, what did the recalculation reveal about planning in changing environments?

Grid Size

Width Height

Theme

Name Your AI

Mode

Heuristic

Animation Speed

FastSlow

A* (A-Star) Algorithm A* finds the shortest path by combining g(n) — actual cost from start — with h(n) — a heuristic estimate to the goal. It always expands the node with the lowest f(n) = g(n) + h(n).

Manhattan Heuristic Counts horizontal + vertical steps to the goal. Best for grid movement with no diagonals. Typically expands fewer nodes than Euclidean on axis-aligned grids.

Euclidean Heuristic Straight-line distance to the goal. More accurate but can over-estimate on grids — may expand slightly more nodes in practice.

Zero Heuristic (Dijkstra) Setting h(n) = 0 turns A* into Dijkstra's algorithm: guaranteed optimal but explores many more nodes since it has no directional guidance.

Fuel Critical Mode Adds a maximum cost budget. If no path exists within the fuel limit, the mission fails — showing how real-world constraints change planning.

Step Mode Advance one node expansion at a time. Watch how the open set (frontier) and closed set (visited) grow — this is the core of how A* "thinks."