OPTIFLOW 3D

• 

  2D top-down view of spatial block placement across Bay 1, Bay 2, and Bay 3 under strict crane constraints.

• 

  OptiFlow 3D algorithm execution: 15/15 blocks successfully planned with 0 violations in 5.0 seconds.

• 

  CLI command preparation in VS Code to execute the scheduling script on a complex JSON data instance

• 

  VS Code terminal showcasing the tail end of the generated JSON operations array for block exits.

• 

  Gantt chart visualization plotting the optimized ENTRY to EXIT timeline for each block across the bays.

• 

  VS Code terminal displaying the structured JSON payload for initial block entry operations.

Inspiration I

The shipbuilding industry faces highly complex planning challenges where both space utilization and scheduling efficiency directly impact productivity and operational costs. We were inspired by the opportunity to solve a real-world industrial optimization problem that combines geometry, logistics, and scheduling. OGC 2026 challenged us to design a system capable of making intelligent placement and timing decisions under strict operational constraints.

What it does.

OPTIFLOW 3D is a spatio-temporal optimization engine that automatically determines where and when shipyard blocks should be placed, processed, and removed. The system optimizes bay utilization, reduces delivery delays, respects operational constraints, and generates valid production schedules while handling complex geometric layouts and scheduling requirements.

How we built it.

We developed OPTIFLOW 3D in Python using a modular optimization architecture. The system starts with an Earliest Due Date (EDD) scheduling heuristic to generate an initial feasible solution. It then applies geometric validation and optimization techniques such as Large Neighborhood Search (LNS) to improve placements, bay assignments, and schedules. A final validation layer checks that all spatial and temporal constraints are satisfied before generating the solution.

Challenges we ran into. Balancing spatial packing and temporal scheduling was our biggest challenge. Improvements in space utilization often created scheduling conflicts, while reducing delays could decrease storage efficiency. Designing an optimization strategy capable of exploring a large solution space within strict execution time limits required extensive experimentation and refinement.

Accomplishments that we're proud of. We successfully developed a complete optimization pipeline capable of generating valid solutions while handling both geometric and operational constraints. We are particularly proud of integrating scheduling, placement, and validation into a unified system that can efficiently tackle large-scale industrial optimization problems.

What we learned. This project strengthened our understanding of combinatorial optimization, industrial scheduling, geometric reasoning, and metaheuristic search techniques. We also learned how to design scalable optimization systems capable of solving real-world logistics challenges under practical constraints.

What's next for OPTIFLOW 3D. Future improvements include more advanced packing strategies, adaptive optimization parameters, additional neighborhood operators for LNS, and hybrid optimization methods to further improve solution quality and computational efficiency. Our long-term vision is to extend OPTIFLOW 3D to broader industrial logistics and planning applications.

Built With

• json
• jupyter
• kymograph
• matplotlib
• module
• optimal
• python
• scipy
• shapely