MCP Optimizer

Mathematical Optimization MCP Server with PuLP and OR-Tools support

python

3

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "dmitryanchikov-mcp-optimizer": {
      "command": "uvx",
      "args": [
        "mcp-optimizer",
        "--transport",
        "stdio"
      ],
      "env": {
        "SERVER_HOST": "0.0.0.0",
        "SERVER_PORT": "8000",
        "TRANSPORT_MODE": "stdio"
      }
    }
  }
}

You run an MCP server that provides optimized solutions across multiple problem types, including linear programming, assignment, knapsack, routing, scheduling, and more. This guide shows practical steps to use the server with MCP clients, how to install it on your system, and important configuration notes to help you get productive quickly.

How to use

You connect to the MCP server from your MCP client using the transport method that best fits your setup. For local development and direct integration, STDIO is the default transport, meaning you interact with the server through standard input and output. For remote or web-based clients, you can use SSE to receive server-sent events over HTTP.

How to install

Prerequisites you should have before installing are Python, a supported package manager, and a container runtime if you plan to use Docker. Pick one installation path and follow the steps.

# Prerequisites (choose one path below)
# 1) Docker (recommended for stability):
# Ensure Docker is installed and running

# 2) Python with pip and venv:
# Install Python from python.org and ensure python3 and pip3 are in your PATH

# 3) uvx (for quick testing):
# Install uvx if you use a local development flow

Choose one installation path and run the following commands.

# Docker (recommended method)
docker run --rm -i ghcr.io/dmitryanchikov/mcp-optimizer:latest

# Python + venv (cross-platform)
python -m venv .venv
source .venv/bin/activate  # Linux/macOS
# or .venv\Scripts\activate  # Windows
pip install mcp-optimizer

# Run in STDIO mode (default)
mcp-optimizer --transport stdio

# Or run via uvx for STDIO
uvx mcp-optimizer --transport stdio

If you are on macOS and want full OR-Tools support, you can install OR-Tools via Homebrew and then install the stable package variant.

# macOS: install OR-Tools then install the stable package
brew install or-tools
pip install "mcp-optimizer[stable]"

Additional notes

Transport options for production and development are described here. Use Docker for production deployments, and use STDIO for direct client integrations. SSE is suitable for remote MCP clients and web-based integrations.

You can also run the server via different entry points shown in examples, such as using uvx for quick tests or running the Python module directly in STDIO mode.

Available tools

linear_programming

Solve linear programming problems with a linear objective and linear constraints using PuLP as the solver backend.

assignment

Solve assignment problems where a set of workers is matched to tasks with the goal of minimizing total cost.

knapsack

Solve knapsack variants to select items that maximize value under capacity constraints.

routing

Solve routing problems such as TSP and VRP with possible time windows.

scheduling

Optimize job or resource scheduling across periods or shifts.

financial

Perform portfolio optimization and financial risk management tasks.

production

Plan production across multiple periods and resources.