A complex big-picture problem can be studied by combining the knowledge of various disciplines' experts through interdisciplinary collaboration. This collaboration carries a considerable challenge, as it needs to handle the interaction of domains without losing their capabilities. One solution for this problem is co-simulation, which creates a virtual representation that can be tested, validated, and improved. Co-simulation provides the chance to get high-level insights into the complex system; optimization takes advantage of this opportunity by exploring the various system configurations to determine the best performance according to a predefined objective. Numerous fields have already employed this combination of co-simulation and optimization in the literature to study and enhance the performance of complex systems. This survey compiles, assesses, and links existing papers on co-simulation-based optimization to give foundations and guidelines for future research on this combined approach. An overview of co-simulation-based optimization from the conceptual, multidisciplinary, and optimization perspectives is presented. To provide general descriptions of the integration process of co-simulation-based optimization, this overview discusses the trends in the literature, which primarily involves a wide range of domains, numerous optimization methods, diverse strategies used to handle interoperability between co-simulation and optimization, and the benefits of generalization tools and frameworks.