Skip to main content
Back to challenges
Challenge

Multi-Agent Power Grid Optimizer

The 'AI boom' is pushing data center developers to innovate power solutions, turning to alternative energy sources amidst grid access challenges. This challenge focuses on creating a sophisticated multi-agent system that optimizes power resource allocation for data centers. Using Claude Opus 4.5 within a LangGraph workflow, developers will design agents that intelligently manage various energy sources (grid, turbines, generators), forecast demand, and minimize operational costs while ensuring uninterrupted service. The system must adapt to fluctuating energy prices and supply constraints, leveraging agent-to-agent communication and MCP tool integration for real-time infrastructure control.

Workflow AutomationHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Design a graph-based multi-agent system using LangGraph and Claude Opus 4.5 for optimizing data center power resource allocation and planning, incorporating MCP tools and A2A protocol.

Delivery guide

How work is evaluated

Evaluation

Evaluation will assess the multi-agent system's ability to minimize operational costs, maximize power uptime, and efficiently switch between various power sources under dynamic simulated conditions. The effectiveness of LangGraph workflows, A2A communication, and MCP tool usage will be key metrics.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Learning goals

What you should walk away with

  • Design MCP-enabled tool integration for controlling simulated energy infrastructure.

  • Develop hybrid instant/deep reasoning systems with Claude Opus 4.1 for dynamic adaptation.

  • Implement the A2A protocol for seamless coordination between diverse power agents.

  • Orchestrate graph-based agent workflows using LangGraph for complex resource management.

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 2 tasks

Challenge setup

Evaluation will assess the multi-agent system's ability to minimize operational costs, maximize power uptime, and efficiently switch between various power sources under dynamic simulated conditions. The effectiveness...

Challenge input

{'scenario_id': 'string', 'grid_price_history': [float], 'generator_cost_per_kwh': float, 'grid_availability_forecast': [bool], 'data_center_demand...

Agent execution

No agent tool is listed for this challenge.

Evaluated output

{'power_plan': [{'time_step': int, 'source': 'grid'|'generator'|'hybrid', 'power_kwh': int}], 'total_cost': float, 'uptime_percentage': float}

Evaluation checks
  • The power plan meets 100% of the data center's demand throughout the scenario.
  • Evidence of simulated MCP tool interactions (e.g., 'grid_status_check', 'generator_activate') within the ag...
Configured evaluation evidence
  • CostEfficiency target: 95
Configured runtime
  • 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

  • The power plan meets 100% of the data center's demand throughout the scenario.
  • Evidence of simulated MCP tool interactions (e.g., 'grid_status_check', 'generator_activate') within the agent's decision log.

Metrics

  • CostEfficiency · Target 95
  • PlanAdaptability · Target 4

Workflow tasks

  1. Task 1

    Input: {'scenario_id': 'string', 'grid_price_history': [float], 'generator_cost_per_kwh': float, 'grid_availability_forecast': [bool], 'data_center_demand_forecast': [int], 'alternative_sources_capacity': int}

    Output: {'power_plan': [{'time_step': int, 'source': 'grid'|'generator'|'hybrid', 'power_kwh': int}], 'total_cost': float, 'uptime_percentage': float}

    Execution harness: Python

  2. Task 2

    Input: {'current_state': {'time_step': int, 'grid_available': bool, 'current_demand': int, 'current_source': 'grid'|'generator'}, 'outage_duration_steps': int, 'generator_capacity': int, 'generator_cost_per_kwh': float, 'data_center_demand_forecast_remaining': [int]}

    Output: {'contingency_plan': [{'time_step': int, 'source': 'generator', 'power_kwh': int}], 'estimated_outage_cost': float}

    Execution harness: Python

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Multi-Agent Power Grid Optimizer