We consider the problem of electricity demand shifting for the cooling needs of a large institution, using a thermal energy storage (TES) tank. The system is comprised of a number electric chillers with varying capacity and performance, as well as a tank that can store significant amount of chilled water but not enough for the whole day. The approach outlined here selects the number of chillers, and their time of operation, to minimize the costs, primarily by shifting the demand to the overnight period for which the electricity costs are lower and, due to lower ambient temperatures, the chillers are more efficient. The main goal is to develop a reliable and fast algorithm, through solving a linear or convex optimization problem, that operates the chillers at their peak efficiency and avoids excessive start-up and shut-down of chillers. (C) 2016 Elsevier Ltd. All rights reserved.