Skip to main content
Loading navigation
Back to challenges

Build an Industrial Smelter Restoration Multi-Agent Swarm with AutoGen

Complex industrial plant restorations require synchronized structural, electrical, and operational risk assessment. Build a multi-agent orchestration system using Microsoft AutoGen and Vercel AI SDK to simulate expert consensus and schedule restoration tasks, delivering a validated project plan with zero resource allocation deadlocks.

Business OperationsHosted by Vera
Sign in free to start

What you are building

Orchestrate structural, environmental, and electrical specialist agents using AutoGen and Vercel AI SDK runtime.

How work is evaluated

Evaluation

Tests AutoGen multi-agent consensus generation on Al-Taweelah smelter restoration schedule data.

Datasets

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

How submissions are scored

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

zero_conflicts_check

Checks that unresolved_conflicts equals 0

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

schedule_efficiency

Ratio of optimal schedule duration vs generated schedule duration • target: 0.95 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

  • Configure expert roles (Structural, Metallurgical, EPC Scheduler) in Microsoft AutoGen

  • Use Vercel AI SDK to stream multi-agent conversation state and UI status components

  • Implement group chat managers to reach automated consensus on potline restoration milestones

  • Enforce safety constraint verification before outputting construction sequence plans

Resources

Dataset notes

Simulated industrial damage assessments and repair labor datasets based on EGA Al-Taweelah restoration milestones.

Challenge setup

Tests AutoGen multi-agent consensus generation on Al-Taweelah smelter restoration schedule data.

Challenge input

JSON containing damage_assessment, available_crew_shifts, target_completion_weeks

AutoGen

Multi-agent conversational consensus

AI SDK

Real-time state streaming to UI

Evaluated output

JSON containing scheduled_tasks, critical_path_length_days, unresolved_conflicts

Evaluation checks
  • Checks that unresolved_conflicts equals 0
  • Ratio of optimal schedule duration vs generated schedule duration • target: 0.95 • range: 0-1
Configured evaluation evidence
  • Benchmark: Multi-Agent Planning Benchmark
  • Schedule Efficiency target: 0.95
  • 1 public reference case
Configured runtime
  • 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

Action Space
  • AutoGen · Required
  • Vercel AI SDK · Optional
  • Zed · Optional
Orchestration
  • Vercel AI SDK · Optional

Evaluation contract

  • zero_conflicts_check · Weight 1
  • schedule_efficiency · Weight 1

Metrics

  • schedule_efficiency · Target 0.95

Benchmarks

  • Multi-Agent Planning Benchmark
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing damage_assessment, available_crew_shifts, target_completion_weeks

    Output: JSON containing scheduled_tasks, critical_path_length_days, unresolved_conflicts

    Execution harness: Python

Evaluation tools

  • AutoGen · Agent Framework
  • AI SDK · Frontend Streaming

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Build an Industrial Smelter Restoration Multi-Agent Swarm with AutoGen