Agentic Sports Scheduler
For sports scheduling optimization, this challenge involves building an advanced, graph-based agent using LangGraph to create optimized sports league schedules. The agent will leverage the anticipated advanced reasoning capabilities of GPT-5, employing 'extended thinking' with adaptive reasoning budgets to navigate the combinatorial complexity of scheduling. This involves managing numerous constraints such as team travel, venue availability, broadcast windows, and player rest days. The core of the solution will be a dynamic LangGraph workflow that can iteratively refine schedules. The agent will integrate with specialized tools, including a simulated constraint solver, and demonstrate the ability to adjust its computational effort ('thinking budget') based on the current state of the schedule and the difficulty of resolving conflicts. This project emphasizes complex problem-solving, efficient resource allocation, and advanced LLM orchestration.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a graph-based agent workflow using LangGraph and GPT-5 for sports scheduling optimization, featuring extended thinking and adaptive reasoning budgets.
How work is evaluated
The evaluation module assesses the LangGraph agent system's ability to generate optimal and feasible sports schedules, demonstrating effective use of extended thinking and adaptive reasoning with GPT-5 to satisfy complex constraints.
Challenge datasets
Public datasets, plus the private datasets you have access to.
What you should walk away with
Optimize a multi-dimensional sports scheduling problem while satisfying numerous constraints.
Implement extended thinking techniques with GPT-5 for combinatorial problem-solving.
Design and apply adaptive reasoning budgets based on task complexity and state.
Integrate specialized tools (e.g., constraint solver) into the LangGraph workflow.
Build complex, stateful agent workflows using LangGraph for iterative optimization.
Reference links and supporting material
Sample data for 3 tasks
Challenge setup
The evaluation module assesses the LangGraph agent system's ability to generate optimal and feasible sports schedules, demonstrating effective use of extended thinking and adaptive reasoning with GPT-5 to satisfy comp...
Challenge input
JSON: {"league_name": "string", "teams": ["string"], "venues": [{"name": "string", "capacity": "integer", "location": "string"}], "constraints": ["...
Agent execution
No agent tool is listed for this challenge.
Evaluated output
JSON: {"schedule": [{"week": "integer", "matchups": [{"home_team": "string", "away_team": "string", "venue": "string", "date": "string", "time_slot...
- The LangGraph agent runs without errors and produces a structured schedule and reasoning trace.
- The output JSON strictly adheres to the specified format for schedule and trace log.
- Evidence that the agent successfully called and utilized the simulated constraint solver tool within the wo...
- ConstraintSatisfactionRate target: 90
- 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 LangGraph agent runs without errors and produces a structured schedule and reasoning trace.
- The output JSON strictly adheres to the specified format for schedule and trace log.
- Evidence that the agent successfully called and utilized the simulated constraint solver tool within the workflow.
Metrics
- ConstraintSatisfactionRate · Target 90
- ScheduleEfficiencyScore · Target 75
- ExtendedThinkingEffectiveness · Target 80
- AdaptiveBudgetUtilization · Target 70
Workflow tasks
- Task 1
Input: JSON: {"league_name": "string", "teams": ["string"], "venues": [{"name": "string", "capacity": "integer", "location": "string"}], "constraints": ["string"], "duration_weeks": "integer"}
Output: JSON: {"schedule": [{"week": "integer", "matchups": [{"home_team": "string", "away_team": "string", "venue": "string", "date": "string", "time_slot": "string"}]}], "violated_constraints": [{"constraint": "string", "details": "string"}]}
Execution harness: Python
- Task 2
Input: JSON: (Same as GenerateSchedule output)
Output: JSON: {"constraint_adherence": {"total_constraints": "integer", "satisfied": "integer", "violated": "integer", "violation_details": [{"constraint": "string", "count": "integer", "examples": ["string"]}]}, "score": "integer"}
Execution harness: Python
- Task 3
Input: JSON: {"trace_log": [{"step": "integer", "agent_action": "string", "thought_process": "string", "budget_allocated": "string", "result": "string"}]}
Output: JSON: {"extended_thinking_evidence": "boolean", "adaptive_budget_evidence": "boolean", "trace_quality_score": "integer"}
Execution harness: Python
Recipe freeze marker
No recipe freeze time is recorded.