A GA-PSO Hybrid Algorithm with Enhanced Variable Boundary Handling

I downloaded a GA-PSO hybrid algorithm from online sources and implemented optimization using combined Genetic Algorithm and Particle Swarm Optimization strategies. Through testing, the algorithm showed good efficiency but revealed several limitations. To address these issues, I modified and improved the original implementation. The primary modification involved enhancing the boundary constraint handling - whereas the original algorithm assumed identical upper/lower limits for all design variables (using uniform boundary arrays), I restructured the code to support independent boundaries per variable through dimension-aware constraint functions. This involved modifying the initialization phase to accept variable-specific bounds arrays and updating the velocity/position update functions with conditional boundary checks. Additionally, I fixed several bugs related to boundary violation handling and population regeneration logic, implementing proper constraint enforcement using penalty functions or reinitialization strategies. These enhancements significantly improve the algorithm's stability and practical applicability for real-world optimization problems with heterogeneous design variables.