Cloud data center is becoming the most essential infrastructure for computing services. In effect, the operational cost of a data center is also increasing drastically. To decrease this cost, consolidation of VMs with less degradation of performance is so important. To guarantee the expected Quality of Service (QOS) the important factors to be controlled are performance of the service including timely leverage and overall resource utilization of the data center. In this paper, we tried to investigate how to efficiently utilize resources with reduced SLA violation in a data center. In order to optimize efficiency, VMs ought to be consolidated as tight as possible. To achieve this, an algorithm based on first fit decreasing (FFD) bin packing is designed and implemented. Hence, the algorithm is implemented on the following three approaches to pursue the goal: a)Deterministic Approach, which is mainly based on mean of the individual VMs;b)Stochastic Approach I, which is basically done by treating individual VMs based on their mean and variances and ;c) Stochastic Approach II, which depends on mean and covariance of individual VMs. The results obtained show that consolidating VMs based on mean and variance(stochastic approach I) performed better than the other two approaches for minimizing total percentage of SLA violation and stochastic approach II performed better than the two approaches for minimizing the number of PMs in consolidation.