🧭 Adaptive Pathfinding Visualizer

A systems-level pathfinding benchmarking tool built with Python and Pygame. Paint terrain, run algorithms, insert dynamic obstacles mid-search, and profile performance — all in real time.

Upgraded from PathfindingProject (CIS 730 — AI course), which benchmarked 5 algorithms in Jupyter. This version adds a real-time interactive visualizer, weighted terrain, D* Lite replanning, and an automated performance profiler.

---

📸 Demo

---

✨ Features

Feature	Description
🗺️ Weighted Terrain	Paint mud (cost 5), highway (cost 0.5), or walls
🤖 5 Algorithms	A*, Dijkstra, BFS, DFS, D* Lite
⚡ D* Lite Replanning	Insert obstacles mid-search without restarting
📊 Performance Profiler	Logs nodes visited, time (ms), memory (KB) per run
🗂️ Grid Presets	Load Default, Maze, or Highway configs from JSON
🎨 Live Painting	Click/drag to draw terrain while the app runs

---

🚀 Setup

git clone https://github.com/DeekshithaKalluri/pathfinder-visualizer.git
cd pathfinder-visualizer
python3 -m venv venv
source venv/bin/activate
pip install pygame networkx psutil
python main.py

---

🎮 Controls

Key	Action
1	A* Algorithm
2	Dijkstra
3	BFS
4	DFS
5	D* Lite
SPACE	Run selected algorithm
T	Cycle terrain brush (obstacle → mud → highway → normal)
Click/Drag	Paint terrain on grid
R	Reset grid
D	Insert dynamic obstacle + D* Lite replan
F1	Load Default preset
F2	Load Maze preset
F3	Load Highway preset
P	Print full benchmark report to terminal

---

🏗️ Project Structure

pathfinder_visualizer/
├── main.py                  # Entry point, event loop
├── grid.py                  # Cell, terrain costs, grid logic
├── renderer.py              # Pygame drawing, UI panel
├── algorithms/
│   ├── astar.py             # A* with terrain-weighted heuristic
│   ├── dijkstra.py          # Dijkstra with weighted edges
│   ├── bfs.py               # Breadth-First Search
│   ├── dfs.py               # Depth-First Search
│   └── dstar_lite.py        # D* Lite with dynamic replanning
├── profiler/
│   └── profiler.py          # Nodes, time, memory benchmarking
└── configs/
    ├── default.json          # Open grid preset
    ├── maze.json             # Wall maze preset
    └── highway.json          # Highway + mud regions preset

---

🧠 Algorithm Comparison

Algorithm	Optimal Path?	Terrain-Aware?	Dynamic Replan?
A*	✅	✅	❌
Dijkstra	✅	✅	❌
BFS	❌ (unweighted)	❌	❌
DFS	❌	❌	❌
D* Lite	✅	✅	✅

---

📈 What the Profiler Tracks

Every time you run an algorithm, the profiler logs:

• Nodes visited — how many cells were explored before finding the path
• Time (ms) — wall-clock execution time using perf_counter
• Memory delta (KB) — RSS memory change via psutil

Press P in the app to print a full comparison table across all runs in the terminal.

---

🌍 Terrain Cost Model

Terrain	Color	Move Cost
Normal	Gray	1
Mud	Brown	5
Highway	Yellow	0.5
Obstacle	Black	∞ (impassable)

Algorithms that use cost (A*, Dijkstra, D* Lite) will naturally route around mud and prefer highways. BFS and DFS ignore cost entirely.

---

🔗 Related

• Base project: PathfindingProject — Academic comparative study of 5 pathfinding algorithms (CIS 730, AI course). Includes full written report and statistical benchmarks across randomized grids.

---

📄 License

MIT — see LICENSE

📚 References

• Koenig, S. & Likhachev, M. (2002). D* Lite. AAAI Conference on Artificial Intelligence. — Algorithm basis for the dynamic replanning implementation.