Pathfinding Visualizer
Run Dijkstra, A*, greedy best-first and breadth-first search on the same grid and compare how many cells each explores to find the same shortest path.
Last reviewed by the Radiatus Cloud team
Need this done properly for your business?
Radiatus delivers secure cloud, DevOps & compliance engineering.
They find the same path and do wildly different amounts of work
Dijkstra and A* both return a shortest path, and A* usually examines a fraction of the cells. The difference is the heuristic: A* adds an estimate of the remaining distance to the cost so far, which biases the search towards the goal instead of expanding evenly in all directions. On an open grid the saving is often five to ten times. Watching both explore the same map is the clearest way to see that the algorithms differ in what they explore rather than in what they return.
An admissible heuristic is what keeps A* correct
The estimate must never exceed the true remaining distance. Manhattan distance is admissible on a four-connected grid and inadmissible on an eight-connected one, where diagonal moves make the true distance shorter than the Manhattan estimate. Use the wrong one and A* returns a path quickly that is not the shortest, and nothing warns you. That single condition is the whole boundary between a correct optimisation and a plausible-looking bug.
Greedy search is fast and wrong
Greedy best-first uses only the estimate and ignores the cost already paid, which makes it very fast and gives no guarantee at all. It walks confidently into dead ends and returns whatever path it stumbles onto. It is worth running alongside the others precisely because the path it returns looks reasonable until compared with the actual shortest one, which is how the failure appears in production too.
Related tools
- Data Collection Analysis — Analyze app description to infer data collection.
- Dark Pattern Detector — Scan UX text for manipulative patterns.
- AI Risk Disclosure — Generate disclosure text for AI features.
- Maturity Radar — Generate a radar chart of security maturity.
Frequently Asked Questions
Do Dijkstra and A* return different paths?
No, both return a shortest path. They differ in how many cells they examine to find it, and A* usually examines far fewer.
What makes a heuristic admissible?
It never overestimates the remaining distance. Manhattan distance is admissible with four-way movement and not with eight-way, where diagonals make the true distance shorter.
Why is greedy search unreliable?
It ignores the cost already paid and follows the estimate alone, so it walks into dead ends and returns whatever it finds. The path looks reasonable until compared with the shortest.
What is the difference between BFS and Dijkstra?
On a grid with equal move costs they are the same algorithm. Dijkstra generalises to weighted edges, where BFS would return the fewest moves rather than the cheapest route.
Why does A* sometimes explore as much as Dijkstra?
Because the heuristic is uninformative for that map. In a maze with many walls the straight-line estimate says little about the real distance, and A* degenerates towards Dijkstra.
Privacy & Security
Everything runs in your browser; nothing is uploaded.
How to Use
Choose an algorithm and maze pattern, then search.
Disclaimer: This tool is provided "as is" without warranty of any kind. Results are for educational and utility purposes.