The optimal virtual machine (VM) provisioning is important for improving power efficiency and resource utilization in a cloud service, where the resource pool is basically characterized by high resource dynamics caused by non-linear variation in the availability of processing, memory size, storage capacity, bandwidth and power drawn resulting from the sporadic nature of workload. In this paper, we develop a novel meta-scheduler (PAFA) that makes use of multi-objective discrete firefly algorithm to efficiently obtain a set of non-dominated solutions to detect and track the changing optimal target servers for the virtual machine provisioning, that considers resource dynamics and the power consumption of the servers with different loads or in different sleep states, which are unconsidered to existing methods, that the total resource wastage and total incremental power drawn by the server pool are simultaneously minimum without compromising the performance objectives. The experimental analysis and comparisons with other methods have shown that the proposed algorithm is efficient and promising method in a resource monitor to find the rate of change in the resource pool and to maintain the supply and demand in equilibrium.