Waste-to-Energy Operational Dispatch Optimizer using LangChain and Ludwig
Casablanca waste-to-energy projects face variable feedstock caloric values and complex thermal conversion constraints. Build a LangChain and Ludwig workflow agent that processes daily municipal waste telemetry to predict optimal furnace operating parameters while maintaining a thermal conversion accuracy within 5 percent.
What you are building
The core problem, expected build, and operating context for this challenge.
Implement a hybrid AI agent utilizing LangChain orchestration and Ludwig model training to automate thermal dispatch for municipal waste incinerators.
How work is evaluated
Evaluates furnace parameter dispatch predictions and agent action correctness against real plant baseline operational data.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Safety Threshold Test
Ensures furnace temperature recommendations remain between 850C and 1100C.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Efficiency MAPE
Mean Absolute Percentage Error between predicted conversion efficiency and gold benchmark. • target: 5 • range: 0-100
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Orchestrate waste feedstock predictive modeling using Ludwig and LangChain
Integrate IoT telemetry data into stateful LangGraph execution graphs
Automate thermal furnace operational parameter dispatch in real time
Evaluate operational efficiency gains against baseline combustion standards
Reference links and supporting material
Simulated daily Casablanca municipal solid waste streams including moisture content, caloric values, tonnage, and historical incinerator sensor metrics.
Challenge setup
Evaluates furnace parameter dispatch predictions and agent action correctness against real plant baseline operational data.
Challenge input
JSON object with daily waste composition, moisture level, and target power output.
LangChain
Agent workflow orchestration
Ludwig
Declarative tabular model inference
Evaluated output
JSON object with recommended airflow, furnace target temperature, and estimated thermal conversion efficiency.
- Ensures furnace temperature recommendations remain between 850C and 1100C.
- Mean Absolute Percentage Error between predicted conversion efficiency and gold benchmark. • target: 5 • ra...
- Benchmark: GCC Energy Operational Optimization Benchmark
- Efficiency MAPE target: 5
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
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
- Ludwig · Required
- LangChain · Optional
- All Hands AI · Optional
- LangChain · Optional
Evaluation contract
- Safety Threshold Test · Weight 1
- Efficiency MAPE · Weight 1
Metrics
- Efficiency MAPE · Target 5
Benchmarks
- GCC Energy Operational Optimization Benchmark
- energy
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with daily waste composition, moisture level, and target power output.
Output: JSON object with recommended airflow, furnace target temperature, and estimated thermal conversion efficiency.
Execution harness: Python
Evaluation tools
- LangChain · Agent Framework
- Ludwig · ML Framework
Recipe freeze marker
No recipe freeze time is recorded.