OptimEngine - Operations Scheduling & Routing Solver
Solves Flexible Job Shop Scheduling (FJSP) and Capacitated Vehicle Routing with Time Windows (CVRPTW) using Google OR-Tools. MCP-native for AI agents. 5 tools: optimize_schedule, validate_schedule, optimize_routing, health_check, root.
Overview
⚡ OptimEngine — Operations Intelligence Solver
The first MCP Server for production scheduling, vehicle routing, and bin packing optimization.
An AI-native solver that assigns tasks to machines, deliveries to vehicles, and items to bins optimally using constraint programming. Built for the agentic economy: AI agents discover it, call it, and pay for it — autonomously.
What It Does
OptimEngine solves three families of NP-hard optimization problems that LLMs cannot compute:
1. Scheduling — Flexible Job Shop (FJSP)
Assign tasks to machines optimally with precedence, time windows, machine eligibility, setup times, priorities, and multiple objectives.
2. Routing — CVRPTW
Assign delivery locations to vehicles optimally with capacity constraints, time windows, service times, GPS coordinates, and multiple objectives.
3. Bin Packing
Assign items to bins/containers optimally with weight/volume constraints, item quantities, group constraints, and multiple objectives.
The core insight: LLMs understand optimization requests in natural language but cannot compute optimal solutions. These are NP-hard problems that require specialized solvers. OptimEngine is that solver, exposed as MCP tools that any AI agent can call.
MCP Tools
| Tool | Problem | Input | Output |
|---|---|---|---|
optimize_schedule | Flexible Job Shop Scheduling | Jobs, tasks, machines, constraints | Optimal schedule + Gantt + metrics |
validate_schedule | Schedule verification | Schedule + constraints | Violations + suggestions |
optimize_routing | Vehicle Routing + Time Windows | Depot, locations, vehicles, capacity | Optimal routes + stop times + metrics |
optimize_packing | Bin Packing | Items (weight/volume), bins (capacity) | Optimal assignments + bin summaries + metrics |
Scheduling Capabilities
| Feature | Details |
|---|---|
| Flexible Job Shop | Tasks can run on multiple eligible machines |
| Precedence | Tasks within a job execute in defined order |
| Time Windows | Earliest start, latest end per job |
| Machine Availability | Machines have operational windows |
| Setup Times | Per-task setup time before processing |
| Priorities | Job priority (1-10) for weighted objectives |
| 4 Objectives | Minimize makespan, total/max tardiness, balance load |
| Schedule Validation | Verify existing schedules, get violation reports |
| Gantt Data | Ready-to-render visualization in every response |
Routing Capabilities
| Feature | Details |
|---|---|
| Capacity Constraints | Per-vehicle maximum load |
| Time Windows | Earliest/latest arrival per location |
| Service Times | Time spent at each delivery point |
| GPS Coordinates | Haversine distance from lat/lon |
| Custom Distance Matrix | Override with your own distances/times |
| Drop Visits | Skip infeasible locations with penalty |
| Per-Vehicle Limits | Max travel time/distance per vehicle |
| 4 Objectives | Minimize distance, time, vehicles, or balance routes |
Packing Capabilities
| Feature | Details |
|---|---|
| Weight + Volume | Dual-dimension capacity constraints |
| Item Quantities | Pack N copies of an item type |
| Bin Types | Multiple bin sizes with different costs |
| Group Constraints | Keep related items in the same bin |
| Max Items per Bin | Limit number of items per container |
| Partial Packing | Allow unpacked items for over-constrained problems |
| 4 Objectives | Minimize bins, maximize value/items, balance load |
Quick Start
1. Install & Run
git clone https://github.com/MicheleCampi/optim-engine.git
cd optim-engine
pip install -r requirements.txt
uvicorn api.server:app --host 0.0.0.0 --port 8000
Server starts at http://localhost:8000. Docs at /docs. MCP endpoint at /mcp.
2. Connect via MCP (Claude Desktop, Cursor, etc.)
{
"mcpServers": {
"optim-engine": {
"command": "mcp-proxy",
"args": ["https://optim-engine-production.up.railway.app/mcp"]
}
}
}
Example — Scheduling
curl -X POST https://optim-engine-production.up.railway.app/optimize_schedule \
-H "Content-Type: application/json" \
-d '{
"jobs": [
{"job_id": "J1", "tasks": [
{"task_id": "cut", "duration": 3, "eligible_machines": ["M1", "M2"]},
{"task_id": "weld", "duration": 2, "eligible_machines": ["M2"]}
], "due_date": 10},
{"job_id": "J2", "tasks": [
{"task_id": "cut", "duration": 4, "eligible_machines": ["M1"]},
{"task_id": "weld", "duration": 3, "eligible_machines": ["M2"]}
], "due_date": 12}
],
"machines": [{"machine_id": "M1"}, {"machine_id": "M2"}],
"objective": "minimize_makespan"
}'
Example — Routing
curl -X POST https://optim-engine-production.up.railway.app/optimize_routing \
-H "Content-Type: application/json" \
-d '{
"depot_id": "warehouse",
"locations": [
{"location_id": "warehouse", "demand": 0},
{"location_id": "customer_A", "demand": 20, "time_window_end": 3000, "service_time": 10},
{"location_id": "customer_B", "demand": 15, "time_window_end": 4000, "service_time": 10},
{"location_id": "customer_C", "demand": 25, "time_window_end": 5000, "service_time": 15}
],
"vehicles": [
{"vehicle_id": "truck_1", "capacity": 40},
{"vehicle_id": "truck_2", "capacity": 40}
],
"distance_matrix": [
{"from_id": "warehouse", "to_id": "customer_A", "distance": 500, "travel_time": 500},
{"from_id": "warehouse", "to_id": "customer_B", "distance": 800, "travel_time": 800},
{"from_id": "warehouse", "to_id": "customer_C", "distance": 600, "travel_time": 600},
{"from_id": "customer_A", "to_id": "warehouse", "distance": 500, "travel_time": 500},
{"from_id": "customer_A", "to_id": "customer_B", "distance": 400, "travel_time": 400},
{"from_id": "customer_A", "to_id": "customer_C", "distance": 700, "travel_time": 700},
{"from_id": "customer_B", "to_id": "warehouse", "distance": 800, "travel_time": 800},
{"from_id": "customer_B", "to_id": "customer_A", "distance": 400, "travel_time": 400},
{"from_id": "customer_B", "to_id": "customer_C", "distance": 300, "travel_time": 300},
{"from_id": "customer_C", "to_id": "warehouse", "distance": 600, "travel_time": 600},
{"from_id": "customer_C", "to_id": "customer_A", "distance": 700, "travel_time": 700},
{"from_id": "customer_C", "to_id": "customer_B", "distance": 300, "travel_time": 300}
],
"objective": "minimize_total_distance"
}'
Example — Bin Packing
curl -X POST https://optim-engine-production.up.railway.app/optimize_packing \
-H "Content-Type: application/json" \
-d '{
"items": [
{"item_id": "laptop", "weight": 3, "volume": 8, "value": 1200, "quantity": 10},
{"item_id": "monitor", "weight": 8, "volume": 25, "value": 500, "quantity": 5},
{"item_id": "keyboard", "weight": 1, "volume": 3, "value": 80, "quantity": 20}
],
"bins": [
{"bin_id": "small_box", "weight_capacity": 20, "volume_capacity": 50, "cost": 5, "quantity": 5},
{"bin_id": "large_box", "weight_capacity": 50, "volume_capacity": 120, "cost": 12, "quantity": 3}
],
"objective": "minimize_bins"
}'
Use Cases
- Manufacturing: Production scheduling for contract manufacturing (cosmetics, pharma, food)
- Logistics: Last-mile delivery routing with time windows and capacity
- Warehouse: Bin packing for palletization, container loading, order fulfillment
- Cloud/IT: Resource allocation (VMs to servers, jobs to clusters)
- Food Delivery: Multi-driver route optimization
- Supply Chain: End-to-end scheduling + routing + packing
Architecture
AI Agent (Claude, GPT, Gemini, etc.)
│
▼ MCP Protocol
┌────────────────────────────────────────┐
│ FastAPI + fastapi-mcp │ ← API layer
├────────────┬────────────┬──────────────┤
│ Scheduling │ Routing │ Bin Packing │
│ CP-SAT │ Routing │ CP-SAT │
│ │ Library │ │ ← OR-Tools solvers
├────────────┴────────────┴──────────────┤
│ Pydantic Models │ ← Schema contract
└────────────────────────────────────────┘
Stack: Python 3.12 · FastAPI · OR-Tools (CP-SAT + Routing) · fastapi-mcp · Pydantic v2
Tests
pip install pytest
python -m pytest tests/ -v
97 tests covering: flexible job shop, time windows, due dates, machine availability, setup times, CVRPTW routing, capacity, GPS distances, bin packing, weight/volume constraints, group constraints, partial packing, and realistic manufacturing/delivery/warehouse scenarios.
Landing Page
Marketplace Listings
- MCPize — 6 MCP tools
- Apify Store — 130k+ users/month
- LobeHub — Top MCP directory
License
MIT
Built with Google OR-Tools — the optimization toolkit used by Google for fleet routing, scheduling, and resource allocation at scale.
Project Info
Created At
3 months agoUpdated At
3 months agoAuthor Name
MicheleCampiStar
-Language
-License
-Recommend Servers
View AllAirtreks Mcp
@SEKeener
a day ago
Fabric
@fabric-agent-plugins
26 minutes ago
Memory
@modelcontextprotocol
a year ago
Mcp Server Chatsum
@chatmcp
summarize chat message
typescript
a year ago