Autonomous Delivery Agent

This project simulates an autonomous delivery agent navigating a 2D grid world. The agent must move from a start location to a goal while considering terrain costs, static obstacles, and dynamic moving obstacles. Different search algorithms are implemented and compared in terms of path cost, nodes expanded, and execution time.

The code file is named delivery_agent.py.

Features

• GridWorld Environment
  • Walls (#)
  • Terrain costs (1-9)
  • Moving obstacles (cyclic paths)
• Algorithms Implemented
  • Breadth-First Search (BFS)
  • Uniform Cost Search (UCS)
  • A* Search
  • Simulated Annealing with Replanning
• Experiment Logging
  • Results saved in results/results.csv
  • Performance plots generated for cost, nodes expanded, and execution time
• ASCII Visualization
  • Prints the grid map with path, start (S), and goal (G)

Installation

Make sure you have Python 3.8+ installed. Install the required libraries: pip install matplotlib pandas

Usage

Run the program with command-line arguments: python delivery_agent.py [OPTIONS]

Options

• --algo Choose which algorithm to run. Values: bfs, ucs, astar, simulated_annealing_replan, all Default: all

• --map Choose which map to run on. Values: small, medium, large, dynamic, all Default: all

• --plot Generate plots from results/results.csv.

Examples

Run all algorithms on all maps: python delivery_agent.py Run A* on the medium map: python delivery_agent.py --algo astar --map medium Run BFS on the dynamic map with moving obstacles: python delivery_agent.py --algo bfs --map dynamic Generate comparison plots after running experiments: python delivery_agent.py --plot

Output

• Console shows algorithm performance: --- small ---

bfs: cost=6, nodes=10, time=0.0001s ucs: cost=6, nodes=8, time=0.0001s astar: cost=6, nodes=6, time=0.0000s simulated_annealing_replan: cost=6, nodes=12, time=0.0012s

• Path visualization in ASCII grid:

S*** 1#1 111 111G

• Results stored in:
  • results/results.csv
  • results/cost_comparison.png
  • results/nodes_comparison.png
  • results/time_comparison.png

Future Improvement

• Support diagonal movement
• Add probabilistic moving obstacles
• Integrate real map inputs from external files
• GUI visualization