TSP MCP Server

TSP MCP Server

A Model Context Protocol (MCP) server for solving the Traveling Salesman Problem (TSP). This server provides optimized algorithms to find the shortest route through a set of cities, making it easy to solve TSP problems through natural language interactions with Claude Desktop.

Features

• 🚀 Multiple Algorithms: Dynamic Programming for small instances (≤10 cities), Nearest Neighbor + 2-opt heuristic for larger instances
• 🎯 Optimal Solutions: Guarantees optimal solutions for small problems, high-quality solutions for larger ones
• 🏙️ Named Cities: Support for both coordinate-only and named city inputs
• 📏 Distance Calculation: Calculate distances for custom routes
• 🔄 Multiple Starting Points: Tests multiple starting points to find better solutions
• 📊 Visual Representations: Generate beautiful SVG visualizations of TSP routes and solutions
• 🎨 Customizable Visualizations: Adjustable canvas size, colors, labels, and styling options

Installation

Prerequisites

• Node.js (version 18 or higher)
• npm or yarn
• Claude Desktop application

Method 1: Install from npm (Coming Soon)

npm install -g tsp-mcp-server

Method 2: Install from GitHub

1. Clone this repository:

git clone https://github.com/yourusername/tsp-mcp-server.git
cd tsp-mcp-server

2. Install dependencies:

npm install

3. Build the project:

npm run build

4. Link globally (optional):

npm link

Claude Desktop Integration

To use this MCP server with Claude Desktop, add the following configuration to your Claude Desktop settings:

On macOS

Edit ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "tsp-mcp-server": {
      "command": "/usr/local/bin/node",
      "args": ["/absolute/path/to/tsp-mcp-server/dist/index.js"],
      "cwd": "/absolute/path/to/tsp-mcp-server",
      "env": {
        "PATH": "/Users/yourusername/Library/Python/3.8/bin:/opt/homebrew/bin:/opt/homebrew/sbin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
      }
    }
  }
}

Important: Replace /absolute/path/to/tsp-mcp-server with the actual path where you cloned the repository (e.g., /Users/yourusername/Cursor/tsp-mcp-server).

On Windows

Edit %APPDATA%/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "tsp-mcp-server": {
      "command": "node",
      "args": ["C:\\absolute\\path\\to\\tsp-mcp-server\\dist\\index.js"],
      "cwd": "C:\\absolute\\path\\to\\tsp-mcp-server"
    }
  }
}

Testing the Connection

After updating the configuration:

1. Restart Claude Desktop completely (quit and reopen)
2. Test the connection in a new conversation:

   Can you solve this TSP problem: cities at (0,0), (1,1), (2,0)?
   

3. Check for errors in Claude Desktop's developer console if the server doesn't respond

Troubleshooting

• Server not found: Verify the absolute paths in your configuration
• Node.js not found: Use the full path to node (/usr/local/bin/node on macOS)
• Permission issues: Make sure the dist/index.js file is executable
• Build not updated: Run npm run build after making changes to the source code

Usage Examples

Once integrated with Claude Desktop, you can use natural language to solve TSP problems:

Basic TSP Problem

"I have 5 cities at these coordinates: (0,0), (1,2), (3,1), (2,4), (4,3). 
What's the shortest route to visit all cities and return to the starting point?"

Named Cities

"Find the optimal delivery route for these locations:
- Seoul: (37.5665, 126.9780)
- Busan: (35.1796, 129.0756)  
- Incheon: (37.4563, 126.7052)
- Daegu: (35.8714, 128.6014)
- Daejeon: (36.3504, 127.3845)"

Route Distance Calculation

"Calculate the total distance if I visit cities in this order: 0 → 2 → 1 → 3 → 4
Cities are at: (0,0), (1,2), (3,1), (2,4), (4,3)"

🎨 Visual TSP Solutions

"Solve the TSP for these 4 cities and create an image visualization: 
(0,0), (5,3), (2,7), (8,1)"

"Create a dark-themed visualization of this route through Seoul, Busan, and Incheon"

"Show me a 1000x800 pixel colorful visualization of the optimal route through these delivery points"

Available Tools

The server provides five powerful tools:

1. solve_tsp

Solves TSP for cities with x,y coordinates.

Input:

{
  "cities": [
    {"x": 0, "y": 0},
    {"x": 1, "y": 2},
    {"x": 3, "y": 1}
  ]
}

2. solve_tsp_with_names

Solves TSP for named cities with coordinates.

Input:

{
  "cities": [
    {"name": "Seoul", "x": 37.5665, "y": 126.9780},
    {"name": "Busan", "x": 35.1796, "y": 129.0756},
    {"name": "Incheon", "x": 37.4563, "y": 126.7052}
  ]
}

3. calculate_route_distance

Calculates the total distance for a specific route.

Input:

{
  "cities": [
    {"x": 0, "y": 0},
    {"x": 1, "y": 2},
    {"x": 3, "y": 1}
  ],
  "route": [0, 2, 1]
}

4. 🖼️ visualize_tsp_route

Generates a high-quality PNG image visualization of a TSP route.

Features:

• 🔴 Red circles for cities with index numbers
• 🔵 Blue lines for route connections
• 📝 City labels with visit order
• 📏 Total distance display
• 🎨 Multiple style themes
• ⚡ Instant image file generation

Input:

{
  "cities": [
    {"x": 0, "y": 0, "name": "Start"},
    {"x": 5, "y": 3, "name": "Middle"},
    {"x": 2, "y": 7, "name": "End"}
  ],
  "route": [0, 1, 2],
  "options": {
    "width": 1000,
    "height": 800,
    "showLabels": true,
    "showDistance": true,
    "style": "dark"
  }
}

5. 🚀 solve_and_visualize_tsp

Solves TSP and generates a beautiful image visualization in one step - the ultimate TSP tool!

Features:

• ⚡ One-step solution and visualization
• 🧠 Automatic algorithm selection (DP vs heuristic)
• 📊 Detailed solution breakdown
• 🖼️ Professional PNG image generation
• 📋 Comprehensive route information
• 🎨 Customizable styling themes

Input:

{
  "cities": [
    {"x": 0, "y": 0, "name": "Warehouse"},
    {"x": 5, "y": 3, "name": "Store A"},
    {"x": 2, "y": 7, "name": "Store B"},
    {"x": 8, "y": 1, "name": "Store C"}
  ],
  "options": {
    "width": 800,
    "height": 600,
    "showLabels": true,
    "showDistance": true,
    "style": "modern"
  }
}

Algorithms

Dynamic Programming (Held-Karp)

• Used for: Problems with ≤10 cities
• Time Complexity: O(n²2ⁿ)
• Space Complexity: O(n2ⁿ)
• Guarantees: Optimal solution

Nearest Neighbor + 2-opt

• Used for: Problems with >10 cities
• Time Complexity: O(n³) for 2-opt improvement
• Approach:
  1. Try multiple starting points with Nearest Neighbor
  2. Improve each solution with 2-opt local search
  3. Return the best solution found

Development

Running in Development Mode

npm run dev

Testing the Server

You can test the server directly:

echo '{"jsonrpc": "2.0", "id": 1, "method": "tools/list"}' | node dist/index.js

Building

npm run build

Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Built with the Model Context Protocol SDK
• Implements classic TSP algorithms adapted for practical use
• Inspired by the need for accessible optimization tools

🎨 TSP MCP 서버의 시각화

TSP MCP 서버는 고품질 PNG 이미지 시각화를 제공합니다:

🖼️ 이미지 기반 시각화

• 기술: HTML5 Canvas + Node.js Canvas 라이브러리
• 구현: ImageTSPVisualizer 클래스 (src/image-visualizer.ts)
• 출력: PNG 이미지 파일로 저장 후 자동 열기
• 장점:
  • 고품질 래스터 이미지
  • 다양한 스타일 테마 지원
  • 글로우 효과 등 고급 그래픽 효과
  • 외부 의존성 없이 독립 실행
  • Claude artifacts 없이 즉시 시스템 파일 생성

지원하는 스타일 테마:

'modern'   // 기본 - 흰 배경, 빨간 도시, 파란 경로
'dark'     // 어두운 테마 - 검은 배경, 밝은 색상
'minimal'  // 미니멀 - 회색톤 단순한 디자인
'colorful' // 컬러풀 - 다채로운 색상 조합

🎯 시각화 사용 방법

TSP 해결 + 시각화 (권장)

{
  "name": "solve_and_visualize_tsp",
  "arguments": {
    "cities": [{"x": 0, "y": 0}, {"x": 5, "y": 3}],
    "options": {
      "style": "dark",
      "width": 1200,
      "height": 800
    }
  }
}

기존 경로 시각화

{
  "name": "visualize_tsp_route", 
  "arguments": {
    "cities": [{"x": 0, "y": 0}, {"x": 5, "y": 3}],
    "route": [0, 1],
    "options": {
      "style": "minimal"
    }
  }
}