Theme: This study investigates the Inventory Routing Problem (IRP) within the context of Vendor-Managed Inventory (VMI), a prevalent supply chain practice where suppliers assume responsibility for replenishment. The IRP, a combinatorial problem that has been widely studied for almost 40 years, encompasses three distinct subproblems: delivery scheduling, inventory management, and vehicle routing. Problem: Despite its age, the IRP continues to attract industry and academia attention. The recent 12th DIMACS Implementation Challenge dedicated a track to the IRP, and among the commonly used benchmarks, 401 instances still lack optimal solutions, particularly in the challenging Large instance subset. Hypothesis and Justification: The HGS framework proposed by Vidal et al. (2012) emerged as a prominent tool used successfully by numerous teams in the competition. However, to the best of our knowledge, the HGS framework has not been tested for the IRP. This study proposes a method combining the HGS framework with an efficient local search strategy, namely NSIRP proposed by Diniz et al. (2020), to tackle the IRP. Methodology: We implemented the proposed method and compared its performance to 21 existing methods using the literature benchmarks. Summary of Results: Our approach identified 79 new Best Known Solutions (BKS) out of 1100 instances. If applied under the same rules as the DIMACS competition, our method would have secured the first place. Contributions and Impacts: This work contribute to the ongoing development of IRP methods, offering an efficient and competitive approach that may inspire further research and practical applications in the realm of inventory management and vehicle routing.