Hybrid Heuristic Optimizer

About the Project

Our project was inspired by the real complexity of shipyard logistics. My dad had experience with this type of industrial work, where large blocks, limited space, and strict deadlines all have to be managed carefully. Hearing about how difficult it can be to organize heavy components in a shipyard made me interested in building an algorithm that could solve this kind of problem more efficiently.

We built a hybrid optimization solver for the Grand Shipyard Puzzle. The main goal of the project is to pack shipyard blocks efficiently while also respecting scheduling constraints. Instead of using only one method, our solver combines greedy heuristics, local search, repair strategies, and simulated annealing to improve the final result.

The solver first creates an initial solution by sorting blocks based on factors such as size, deadline, and difficulty of placement. After that, it checks for conflicts and tries to improve the solution step by step. A simplified version of the scoring idea is: Score = Reward - Penalty where penalties can come from overlaps, delays, inefficient space usage, or invalid placements. During this project, we learned how challenging combinatorial optimization can be. A small change in the order of blocks or the placement strategy can create a very different final result. We also learned that a good optimization algorithm needs to balance speed and accuracy, especially when it must run under strict time limits.

The biggest challenge we faced was creating a solver that could find good solutions quickly. Searching for the perfect answer would take too much time, so our approach focuses on always keeping the best solution found so far while continuously trying to improve it.

Overall, this project helped us understand how algorithms can be used to solve real industrial problems. It showed us how packing, scheduling, logistics, and optimization can work together to make complex systems more efficient.

Built With

• batch
• python