Recent years have witnessed the widespread use of highly efficient energy systems as an important measure to reduce not only energy consumption but also operating costs. A ground source heat pump system has been attracting considerable attention because of its high efficiency. Although many studies have been conducted to investigate and evaluate the ground source heat pump's performance, they have not sufficiently studied its optimal operation considering dynamic ground temperature variation caused by the high thermal capacity of the ground. Calgulations considering both thermal history of the ground and optimal load dispatch are complicated and thus entail high computation costs. In this paper, an efficient optimization method is proposed to determine optimal operations of a hybrid ground source heat pump system that is used to handle the cooling load and hot water demand. The proposed method, namely epsilon-constrained differential evolution with random jumping, can solve nearly all possible configurations and is a suitable method for the nonlinear configuration used herein because the ground source heat pump has highly nonlinear characteristics and the ground temperature calculation cannot be simplified to a linear formulation. The optimal operations achieved by the proposed method can reduce operating costs by at least 3.78% and at most 12.56% compared to empirical operations. In addition, the proposed method derives the solution rapidly while maintaining high computation accuracy. Therefore, it can be used in practical situations to determine an optimal operating schedule as a day-ahead optimization. (C) 2017 Elsevier Ltd. All rights reserved.