Visual Simulator for Directed Graph Search Algorithms: Complete Implementation from BFS to A*

This article introduces a teaching-level visual simulator for directed graph search algorithms implemented in Python, supporting seven classic algorithms (BFS, DFS, UCS, DLS, IDDFS, Greedy, A*), with step-by-step visual output, suitable for AI beginners. The project comes from GitHub (nataliettv/AI-search-algorithms), is a final project for an AI course, has no external dependencies, and demonstrates the algorithm execution process through console output, helping learners intuitively understand the principles and differences of different search strategies.

• Original author/maintainer: nataliettv
• Source platform: GitHub
• Project name: AI-search-algorithms
• Original link: https://github.com/nataliettv/AI-search-algorithms
• Release date: May 27, 2026
• Project nature: Final project for an artificial intelligence course This project is designed specifically for AI beginners, requires no external dependency libraries, and helps understand abstract algorithms through detailed comments and step-by-step output.

The project implements seven search algorithms, covering blind to heuristic search:

1. BFS (queue, level exploration, optimal in steps); 2. DFS (recursive backtracking, low memory usage); 3. UCS (priority queue, cost-optimal); 4. DLS (DFS with depth limit); 5. IDDFS (combines BFS's optimality and DFS's low memory); 6. Greedy (heuristic-guided, fast but not optimal); 7. A* (f(n)=g(n)+h(n), optimal under admissible heuristics). Core implementation: Uses Python dictionaries to represent weighted directed graphs, supports heuristic function definition, each algorithm is an independent function with clear structure for easy comparison.

Input Methods:

1. Interactive input: Enter nodes, edges, and heuristic values step-by-step in the console;
2. File input: Supports loading graph definitions (including start/goal nodes, edges, heuristic values) from text files, with preset graphs in the examples directory. Visualization Output: Each step shows the current node, visited set, path, queue/stack status, etc. For example, A* output displays the calculation process of f(n)=g(n)+h(n), helping to understand the role of heuristics. Code examples: Graph structure: grafo = {'A': {'B':1, 'C':4}, ...} Heuristics: heuristicas = {'A':6, 'B':4, ...}

Algorithm Comparison: Allows intuitive comparison of completeness (BFS/UCS/IDDFS/A* are complete), optimality (UCS/A* are optimal), efficiency (Greedy is fast), and memory usage (DFS/IDDFS are memory-friendly). Heuristic Understanding: Adjusting heuristic values allows observing changes in A* performance (more accurate heuristics lead to fewer expanded nodes; zero heuristic degrades to UCS). Code Learning: Each algorithm is an independent function with similar structures for easy comparison, detailed comments (including original Spanish text), suitable for beginners to learn implementation.

Summary: This simulator is an excellent educational tool that converts abstract theory into runnable code and visual output, helping learners deeply understand the essence of algorithms, suitable for AI students, competition participants, self-taught developers, etc. Expansion Directions: Add graphical interfaces (Pygame/Tkinter), more algorithms (bidirectional search, RBFS), automatic testing, performance analysis, practical applications (maze solving, path planning), etc.