This research is focused on finding an assignment that minimizes the number of physical machines (PMs) that are used for mapping a collection of virtual machines (VMs) for the deployment of the cloud-based Advanced Distribution Management System (ADMS). The paper compares two approaches: a) one that is based on the exhaustive search and b) the multiple constraints bounded Knapsack algorithm based approach. An example of optimal mapping for multiple ADMS deployment is used to illustrate the reduction in number of physical resources needed. Also, it compares the required duration of the exhaustive search and the Knapsack algorithm execution, and illustrates the potential resource optimization depending on the overcommitment level. It is shown that virtually enabled resource sharing enables significant reduction of allocated physical resources and that the Knapsack based approach is more efficient (as compared to the exhaustive search) and can potentially lead to a fully dynamic placement and deployment reconfiguration in real-time even for a case of a large number of VMs.