🅿 Smart Parking & Route Planning Visualizer

An AI-powered Java application combining A*, Dijkstra, traffic prediction, and dynamic parking allocation.

---

🚀 How to Run

# Compile (requires JDK 17+)
mkdir -p out
find src -name "*.java" -print0 | xargs -0 javac -d out

# Run
java -cp out src.ui.MainFrame

Or via IDE: Open the project, set src.ui.MainFrame as the main class and run.

---

🎯 Use Case

A driver enters a parking lot. The system:

1. Predicts traffic/congestion on each lane using time-of-day patterns
2. Finds available spots that match the driver type (Regular / Handicapped / EV / VIP)
3. Runs A* or Dijkstra to find the optimal route to the best spot
4. Lets the driver reserve the spot with a token and TTL
5. Replans dynamically as congestion changes (heatmap updates every 5s)

---

🏗 Architecture

src/
├── core/
│   ├── Node.java              — Grid cell with parking metadata + A* costs
│   ├── Grid.java              — Grid logic: walls, spots, reservations, neighbors
│   ├── Heuristic.java         — Manhattan / Euclidean / Chebyshev / Weighted
│   ├── AStarAlgorithm.java    — A* with pluggable heuristic + AI-tuned edge costs
│   ├── DijkstraAlgorithm.java — Dijkstra (optimal, explores more than A*)
│   ├── PathfindingAlgorithm.java — Shared interface
│   └── LocalServer.java       — In-memory reservation manager (token + TTL)
│
├── ai/
│   ├── TrafficPredictor.java  — AI congestion prediction (time-of-day + hotspots)
│   ├── HeuristicTuner.java    — Dynamically adjusts edge costs for congestion avoidance
│   └── ParkingAllocator.java  — Core AI: scores spots, runs pathfinding, returns best result
│
├── ui/
│   ├── GridPanel.java         — Renders grid, heatmap, paths, spots; handles interaction
│   ├── ControlPanel.java      — Algorithm selector, driver type, AI controls
│   └── MainFrame.java         — App entry point (main method here)
│
└── utils/
    ├── ParkingLotLayout.java  — Generates realistic parking lot (lanes + bays + special spots)
    ├── MazeGenerator.java     — Recursive backtracker maze (for algorithm demos)
    └── TimerUtil.java         — High-resolution timing utility

---

🧠 AI Components

TrafficPredictor

• Simulates real-world traffic patterns based on time of day (Morning Rush, Midday, Evening Rush, Night, Weekend)
• Applies spatial hotspots (entrance/exits are busier)
• Uses Gaussian noise for realistic variation
• Smooth blending — weights update gradually every 5 seconds (no jarring jumps)

HeuristicTuner

• Bridges traffic prediction with pathfinding
• Applies congestion penalty to edge costs: tunedCost = baseCost × (1 + scale × congestion)
• Supports inflated heuristic (Weighted A*) for speed/optimality tradeoff

ParkingAllocator

• Scores all available spots: score = distance + congestionPenalty + typeBonus
• Evaluates Top-K candidates to avoid dead-ends
• Filters by driver type preference (e.g., EV driver gets EV spots)
• Returns full result: best spot, path, explored nodes, timing stats

---

🎮 Controls

Action	Effect
Right-click	Set Start → then Goal
Left-click	Toggle wall
Shift + Drag	Draw walls
Right-click on spot	Context menu: Reserve / Info
Space	Run selected algorithm
A	AI auto-allocate best parking spot
H	Toggle traffic heatmap
C	Clear path
R	Reset everything

---

📊 Algorithm Comparison (built-in)

Click "⚖ Compare A* vs Dijkstra" to run both on the same grid:

• Yellow path = A* result
• Purple path = Dijkstra result
• Stats panel shows: nodes explored, path cost, runtime for each

Typically A* explores far fewer nodes than Dijkstra while finding the same (or similar) cost path.

---

🗂 Spot Types

Symbol	Type	Color
P	Regular	Light blue
♿	Handicapped	Blue
⚡	EV Charging	Green
★	VIP	Gold
✖	Occupied	Red
R	Reserved	Orange

---

📦 Dependencies

• Java 17+ (uses switch expressions)
• Swing (built-in, no external libraries needed)