The　emerging　medical　alliance　in　some　countries　enables　the　joint　optimization　for　its　members　through　timely　necessary　information　exchanging　and　sharing.　This　paper　investigates　a　joint　decision-making　optimization　problem　between　a　focal　hospital　and　multiple　pharmaceutical　manufacturers　in　a　medical　alliance,　to　optimize　both　parties＇　objectives　from　the　supply　chain　management　perspective.　Specifically,　the　hospital　makes　the　decision　on　the　medicine　procurement　from　multiple　candidate　pharmaceutical　manufacturers　to　minimize　its　total　procurement　cost,　and　each　chosen　pharmaceutical　manufacturer　determines　the　production　plan　for　the　different　types　of　medicine　procured　by　the　hospital　to　maximize　its　total　profit.　Given　that　the　hospital　and　pharmaceutical　manufacturers　make　decisions　under　their　objective　functions　and　constraints　while　the　decisions　affect　each　other＇s　objective　mutually,　a　bi-level　programming　mathematical　model　is　established,　with　the　hospital　acting　as　the　leader　and　the　pharmaceutical　manufacturers　as　multiple　independent　followers.　We　propose　an　improved　hierarchical　loop　framework,　including　the　upper　solving,　the　lower　solving,　and　the　improvement　module.　Specifically,　we　embed　the　Variable　Neighborhood　Search　(VNS)　algorithm　in　the　lower-solving　module　and　apply　a　set　of　heuristic　rules　in　the　improvement　module.　To　test　the　performance　of　the　proposed　algorithm,　we　replace　the　VNS　algorithm　in　the　lower　solving　module　with　PSO,　GA,　and　SA,　and　compare　their　performance　in　small,　medium,　and　large-scale　instances.　Besides,　we　also　validate　the　effectiveness　of　the　improvement　module　by　comparing　the　performance　of　the　algorithm　with　and　without　the　module.　The　computational　experiments　prove　the　effectiveness　and　efficiency　of　the　proposed　approach.