Multi-Agent Resource Optimization for In-Space Factories
The advent of in-space manufacturing requires sophisticated autonomous systems for resource management, production scheduling, and failure recovery. This challenge focuses on developing an intelligent, multi-agent system to optimize the utilization of limited resources (power, raw materials, robot time) within an in-space manufacturing platform. The system must adapt to dynamic production demands, prioritize tasks, and handle unexpected events such as equipment failures or material shortages, aiming to maximize throughput and operational resilience. Participants will design and implement a simulation of an in-space factory featuring multiple manufacturing units and a shared resource pool. The core of the solution will involve leveraging Mistral Large 2 for high-level decision-making and strategic optimization, orchestrating agents within a multi-agent framework. `DuckDB` will be used for rapid, on-the-fly analytical processing of factory telemetry and resource logs to inform the AI's real-time decisions. The goal is to build a robust system that can efficiently manage production, minimize waste, and ensure the continuous operation of critical manufacturing processes in an autonomous and resilient manner, crucial for future domestic microelectronics production in space.
What you are building
The core problem, expected build, and operating context for this challenge.
Implement a multi-agent system using Mistral Large 2 and DuckDB to optimize resource allocation and production scheduling for an autonomous in-space manufacturing platform.
How work is evaluated
The evaluation will assess the efficiency and resilience of the autonomous resource optimization system within a simulated in-space manufacturing environment. Participants will submit their simulation code and a report detailing the system's performance under various scenarios.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Integrate and utilize DuckDB for real-time analytical processing of operational data to support AI decisions.
Implement multi-agent systems for distributed resource management and task coordination.
Design and simulate an autonomous in-space manufacturing environment.
Orchestrate and manage agent interactions using a multi-agent framework to achieve global objectives.
Leverage large language models (Mistral Large 2) for strategic decision-making in complex operational scenarios.
Reference links and supporting material
Sample data for 2 tasks
Challenge setup
The evaluation will assess the efficiency and resilience of the autonomous resource optimization system within a simulated in-space manufacturing environment. Participants will submit their simulation code and a repor...
Challenge input
JSON object containing simulation parameters (e.g., {'duration_hours': 24, 'initial_resources': {'power': 1000, 'material_A': 500}, 'production_ord...
Agent execution
No agent tool is listed for this challenge.
Evaluated output
JSON object with 'total_units_produced' (integer) and 'resource_utilization' (object, e.g., {'power': 0.85}).
- Ensures the submitted Python simulation code runs without fatal errors across various test scenarios.
- Verifies that the system never over-allocates resources beyond available capacity at any point in the simul...
- AverageThroughput target: 1.5
- Python execution harness
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- Ensures the submitted Python simulation code runs without fatal errors across various test scenarios.
- Verifies that the system never over-allocates resources beyond available capacity at any point in the simulation.
Metrics
- AverageThroughput · Target 1.5
- MeanTimeBetweenFailures · Target 300
- ResourceEfficiency · Target 0.85
Workflow tasks
- Task 1
Input: JSON object containing simulation parameters (e.g., {'duration_hours': 24, 'initial_resources': {'power': 1000, 'material_A': 500}, 'production_orders': [{'type': 'chip_A', 'quantity': 10}]}).
Output: JSON object with 'total_units_produced' (integer) and 'resource_utilization' (object, e.g., {'power': 0.85}).
Execution harness: Python
- Task 2
Input: JSON object containing simulation parameters and a 'failure_event' (e.g., {'time': 12, 'type': 'equipment_failure', 'target_unit': 'unit_3', 'duration_hours': 4}).
Output: JSON object with 'production_loss_percentage' (float) and 'recovery_time_hours' (float).
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.