Metaheuristic algorithms have become indispensable tools for solving complex combinatorial optimization problems. However, their performance often depends critically on the selection of internal parameters, which are frequently tuned in an ad hoc manner. This paper investigates the hierarchical nested structure of the metaheuristic algorithm and its impact on optimization performance, where parameters of one metaheuristic are optimized using another, resulting in a multi-level optimization framework. We demonstrate this concept using a four-tier architecture: the Genetic Algorithm (GA) optimizes the Radius (R) parameter in the Circle Group Heuristic (CGH), which in turn constructs high-quality initial populations for the Discrete Bacterial Memetic Evolutionary Algorithm (DBMEA). The DBMEA itself is a memetic algorithm that integrates global evolutionary mechanisms from Bacterial Evolutionary Algorithm (BEA) and with local search strategies (2-OPT and 3- OPT), thus comprising two inherent levels. Together, this nesting creates a four-level metaheuristic hierarchy. The DBMEA is then applied to solve variants of NP-Hard problems such as Traveling Salesman Problem (TSP). Our experiments on benchmark datasets show that this nested structure not only improves convergence speed and solution quality but also demonstrates the potential of deeply nested metaheuristic designs for scalable, robust optimization.