In dynamic manufacturing contexts, the Job-Shop Scheduling Problem (JSP) presents considerable hurdles, particularly in light of uncertainties such as machine breakdowns and job reworks. In this article, multiple scheduling algorithms are evaluated in various scenarios, with a special emphasis on lowering makespan, or overall completion time. Advanced algorithms like Genetic Algorithm (GA) and Deep Reinforcement Learning (DRL) are contrasted with heuristic techniques like Shortest Processing Time (SPT), Longest Processing Time (LPT), and First-In-First-Out (FIFO). The research emphasizes the advantages of the hybrid Ant Colony Optimization with Genetic Algorithm (ACO_GA) solution that has been suggested. Although GA and DRL produce outcomes that are competitive, ACO_GA regularly performs better than them and, in most cases, comes quite close to the optimal solution. Even in the face of difficult, real-world problems, our hybrid approach yields more reliable and nearly optimal scheduling solutions by fusing the exploration potential of ACO with the refinement power of GA. The results show how adaptable and successful ACO_GA is, which makes it the best option for solving JSP in dynamic contexts.