Hybrid Combinatorial Problems Used for Multimodal Optimisation

This work presents an integrating clustering and local search heuristics for multi-domain optimization applications in the Traveling Salesman and Protein Structure Prediction problems, evaluating their efficacy across diverse optimization scenarios, including single-objective, multicriteria, dynamic, and multimodal tasks. GA framework is designed for solving the Traveling Salesman Problem (TSP), incorporating clustering and 2 -opt local search to refine solutions. The methodologies from the TSP are extended to address the Protein Structure Prediction (PSP) problem using the Hydrophobic-Hydrophilic (HP) model. This novel adaptation showcases the adaptability of GAs in addressing structural prediction in computational biology, defining how route optimization strategies can be transformed to minimize conformational energy and improve folding patterns. Experimental results demonstrate the hybrid approach in achieving both route optimization and stable protein conformations, thereby bridging combinatorial optimization techniques with structural biology applications.