Abstarct: In large water supply systems, several pump stations are usually used to provide pipelines head pressure. Each station may consist of some parallel pumps to provide water pipeline discharge. By sudden failure of these pumps, an unsteady flow will occur which cause pressure fluctuation. Assessment and control the maximum and the minimum pressure of pipeline is essential to design pipeline and utilize it. One method to diminish the harmful effects of water hammer is step by step shut down. The main purpose of this thesis is determining the interval between the steps and the pattern of this shutdown somehow that the maximum pressure reduce and the minimum pressure rise up and stay in a safe range. First, a numerical model of a comprehensive pumping system consists of some stations and diversity of parallel pump in the stations is simulated. This model simulates unsteady flows of sudden failure. The equations of this modeling are flow equations (continuity and momentum) and pump equations (head and torque). This model is recalled as a function to invasive weeds algorithm (IWO). The intervals between two steps of failure are reproduced by standard IWO and failure pattern is reproduced by binary_IWO. The maximum and minimum pressures are optimized by IWO. Finally, the optimized variables are expanded into the different length of pipelines and number of stations. The results illustrated that the shutdown pattern could have a significant effect on the pipeline's pressure. The harmful effect of water hammer can be restricted only by choosing the appropriate pattern and interval of two steps of the shutdown. Since these two variables can have a wide range of diverse values, using an optimizing algorithm will be effective. The results of optimizing are intensively depended on power and number of pumps, length of pipes …, so optimized variables are valid in a specific pumping system. However, a limit range of these variables can be presented for a specific pumping system and the expansion of this optimization also is given for the different length of pipeline and number of station. The result demonstrates optimized variables are also effective for these changes, but by increasing the length of pipe, the maximum pressure rise and the minimum pressure falls down and conversely, by reducing the pipeline's length, the maximum and minimum pressure respectively, decrease and increase.