Simulation-optimization (Sim-Opt) is a widely used optimization technique that enables the use of simulation model so as naturally describe system complexity and stochastics. A key barrier to its broader adoption is the high computational cost associated with simulation that often limits its practicability. In this paper, we propose the use of GPU parallel computing, to enhance the computational efficiency of Sim-Opt. The main objective of this work is to develop a systematic framework that can be used to construct an efficient hybrid CPU-GPU program. The optimization of a process monitoring model using a Genetic Algorithm is used as a case study to illustrate the proposed approach. Our results show an over 100× acceleration of computation time by the developed hybrid program in comparison to a traditional CPU-based approach. © 2016 Elsevier Ltd.