TACAS — Temporal-Aware Crane Accessibility Search
Inspiration
Modern shipyards manage thousands of large ship blocks across limited work bays while operating under strict deadlines and resource constraints. Traditional approaches often focus on packing efficiency or scheduling independently, but real shipyard operations require balancing both spatial and temporal constraints simultaneously.
The Optimization Grand Challenge 2026 inspired us to explore a different perspective: instead of optimizing only where a block can be placed, we focus on whether that block can be efficiently accessed and removed in the future. This shift transforms the problem from simple packing into a long-term planning challenge.
What It Does
TACAS (Temporal-Aware Crane Accessibility Search) is a shipyard optimization framework designed to minimize project delays while maintaining efficient bay utilization and future block accessibility.
The system generates optimized plans for:
- Bay assignment
- Block placement
- Orientation selection
- Entry scheduling
- Exit scheduling
- Crane-access-aware operations
By considering future extraction feasibility during optimization, TACAS aims to reduce congestion, avoid operational deadlocks, and improve overall schedule performance.
How We Built It
Our approach combines multiple optimization techniques:
- Constraint-aware scheduling to prioritize critical blocks and reduce tardiness.
- Geometry-based placement evaluation for efficient bay utilization.
- Crane accessibility analysis to estimate future extraction feasibility.
- Adaptive local search to continuously improve placement and scheduling decisions.
- Multi-objective optimization balancing tardiness, preference scores, and workload distribution.
The framework evaluates both spatial and temporal consequences of each placement decision rather than optimizing only immediate packing efficiency.
Challenges We Faced
The most difficult aspect of the problem is the interaction between space and time.
A placement that appears optimal today may create severe accessibility issues in the future. Similarly, reducing congestion can conflict with due-date requirements and bay utilization objectives.
Designing a system capable of balancing these competing objectives while remaining computationally efficient is the primary challenge addressed by TACAS.
What We Learned
This project highlights how real-world industrial optimization extends far beyond traditional scheduling or packing problems. Effective solutions require understanding future operational consequences, resource accessibility, and long-term system behavior.
The challenge provided valuable insights into operations research, large-scale optimization, industrial logistics, and constraint-aware decision-making.
Future Improvements
Future versions of TACAS will explore:
- Machine-learning-guided heuristic selection
- Congestion prediction models
- Dynamic crane movement optimization
- Large Neighborhood Search (LNS) enhancements
- Hybrid optimization using Operations Research and AI
Our goal is to develop a robust industrial-grade decision-support system capable of handling increasingly complex shipyard environments.
Log in or sign up for Devpost to join the conversation.