An interactive engineering tool designed to visualize, analyze, and compare Pathfinding algorithms on unweighted graphs, procedurally generated using BFS with visited node tracking.
The project allows for real-time transformation of the maze topology (from Tree to Cyclic Graph) to test the robustness of heuristic algorithms.
To empirically analyze the behavior of search algorithms under different topological conditions:
- Perfect Mazes (Trees): Single possible path. Greedy and A* perform similarly.
- Graphs with Cycles (Loops): Multiple paths. Demonstrates how Greedy sacrifices optimality for speed, while A* guarantees the shortest path.
Includes a Real-Time Metrics Panel to measure:
- CPU Time (ms).
- Memory Efficiency (Nodes explored).
- Solution Quality (Path length).
| Algorithm | Structure | Observed Behavior |
|---|---|---|
| BFS (Breadth-First) | Queue (FIFO) | Guarantees optimal path. Uninformed search, highly memory-intensive. |
| DFS (Depth-First) | Stack (LIFO) | Fast but erratic. Frequently produces suboptimal and long solutions. |
| Greedy Best-First | Priority Queue | Very fast. Guided solely by heuristic. In graphs with loops, it often falls into local minima (long paths). |
| A (A-Star)* | Priority Queue | Optimality and efficiency. Guarantees the optimal path while reducing the number of visited nodes. |
R: Generate new maze (Recursive Backtracker).C: Clear the maze.W: Break Walls (Introduces random loops into the graph).B: Run BFS.D: Run DFS.G: Run Greedy.A: Run A*.