Phosphate Rail Track Scheduling Agent with Claude Agents SDK
Saudi Arabia's $1bn phosphate rail track doubling project demands intelligent freight scheduling to maximize throughput without bottlenecking single-line segments. Build an autonomous scheduling agent using Claude Agents SDK with extended thinking to optimize train departure timetables and track utilization. Achieve 95% throughput efficiency without causing block collisions.
What you are building
The core problem, expected build, and operating context for this challenge.
Build a freight train dispatch agent using Claude Agents SDK extended reasoning features.
How work is evaluated
Evaluates rail schedule optimization and collision prevention produced by Claude Agents SDK agent.
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.
Zero Collision Guarantee
Fails if collision_detected is true or block times overlap.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Schedule Optimization Efficiency
Ratio of optimal transit time to total agent-scheduled transit time. • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Build a autonomous dispatch agent using Claude Agents SDK
Utilize extended thinking capabilities to calculate complex train passing schedules
Define custom tool schemas for SAR (Saudi Arabia Railways) track segment blocks
Minimize siding wait times for heavy phosphate transport trains
Reference links and supporting material
Saudi Arabia Railways (SAR) phosphate corridor track topology and freight movement schedules.
Challenge setup
Evaluates rail schedule optimization and collision prevention produced by Claude Agents SDK agent.
Challenge input
JSON object with train_list, single_track_blocks array, double_track_blocks array
Claude Agents SDK
Assigned framework for high-reasoning extended thinking task.
Evaluated output
JSON object with train_schedules array, collision_detected, total_transit_hours
- Fails if collision_detected is true or block times overlap.
- Ratio of optimal transit time to total agent-scheduled transit time. • target: 0.95 • range: 0-1
- Benchmark: LogisticsAgentBench
- Schedule Optimization Efficiency target: 0.95
- 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
- RAI · Required
- Agency (Observability) · Optional
- Synthflow · Optional
Evaluation contract
- Zero Collision Guarantee · Weight 1
- Schedule Optimization Efficiency · Weight 1
Metrics
- Schedule Optimization Efficiency · Target 0.95
Benchmarks
- LogisticsAgentBench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with train_list, single_track_blocks array, double_track_blocks array
Output: JSON object with train_schedules array, collision_detected, total_transit_hours
Execution harness: Python
Evaluation tools
- Claude Agents SDK · Autonomous Agent Framework
Recipe freeze marker
No recipe freeze time is recorded.