Multi-Agent Industrial Asset Management Strategy Engine with CrewAI
Isolated maintenance logs and inventory systems slow down plant downtime responses. Build a multi-agent orchestration team using CrewAI where specialized Maintenance Planner, Inventory Controller, and Production Scheduler agents collaborate to automatically convert predictive maintenance alerts into prioritized work orders.
What you are building
Orchestrate a multi-agent crew using CrewAI to automate predictive maintenance work order creation, spare part reservation, and downtime scheduling.
How work is evaluated
Evaluates the collaborative decision quality of the CrewAI agents when responding to a high-priority bearing vibration failure alert.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
part_availability_check
Ensures scheduled work order only uses available in-stock parts.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
downtime_efficiency_score
Measures how effectively scheduled maintenance minimizes production loss. • target: 0.92 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Configure role-based agents in CrewAI for maintenance, inventory, and scheduling
Implement sequential and hierarchical collaboration workflows between crew members
Integrate synthetic vibration anomaly signals with SAP PM/Maximo work order tools
Optimize production schedule impact while maintaining high overall equipment effectiveness (OEE)
Resources
Vibration time-series data, CMMS spare parts inventory levels, and weekly manufacturing line schedules.
Challenge setup
Evaluates the collaborative decision quality of the CrewAI agents when responding to a high-priority bearing vibration failure alert.
Challenge input
JSON containing machine_id, vibration_mm_s, baseline_threshold, and spare_parts_stock.
CrewAI
Provides built-in collaboration and task delegation mechanics for specialized roles.
Evaluated output
JSON containing finalized_work_order with scheduled_time, part_reserved, and downtime_hours.
- Ensures scheduled work order only uses available in-stock parts.
- Measures how effectively scheduled maintenance minimizes production loss. • target: 0.92 • range: 0-1
- Benchmark: Multi-Agent Maintenance Scheduling Benchmark
- Downtime Efficiency Score target: 0.92
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
- CrewAI · Required
- crewAI · Optional
Evaluation contract
- part_availability_check · Weight 1
- downtime_efficiency_score · Weight 1
Metrics
- downtime_efficiency_score · Target 0.92
Benchmarks
- Multi-Agent Maintenance Scheduling Benchmark
- robotics
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing machine_id, vibration_mm_s, baseline_threshold, and spare_parts_stock.
Output: JSON containing finalized_work_order with scheduled_time, part_reserved, and downtime_hours.
Execution harness: Python
Evaluation tools
- CrewAI · Multi-Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.