The Cloud Computing Load Forecasting Algorithm Based on Kalman Filter and ANFIS

The load forecasting in the cloud computing is one of the most important technologies to ensure the maximize utilization of the system resource. Under the premise that the load is known in the next stage, the cloud computing center can assign the physical machines in advance, thereby reducing the waiting time of the task, and can also reduce the cloud computing center energy consumption. This paper proposed a load forecasting algorithm based on the Kalman filter and adaptive neuro-fuzzy inference system (ANFIS), obtained more accurate load sequence by the kalman filter eliminate observation error, used ANFIS to forecast the load sequence. The predicted results were compared with the original ANFIS algorithm, Autoregressive Integrated Moving Average (ARIMA) algorithm. The K-ANFIS algorithm had improved the prediction accuracy significantly compared with the other two algorithms.