Genetic algorithms (GAs) are a type of search procedure that emulates the process of natural selection and genetics to solve complex optimization problems. The effectiveness of GAs in finding global solutions has led to their widespread applications in solving diverse optimization problems. However, their search performance and convergence not only depend highly on the operators used but are also sensitive to the choice of control parameters. This article presents an improved real-coded GA, called hybrid genetic algorithm (HGA), that employs affine combination-based reproduction and non-uniform mutation. The reproduction is a formula-based operator that helps to improve the convergence and introduce some degree of genetic diversity in the HGA. The non-uniform mutation helps to further maintain diversity within the population and prevent premature convergence to suboptimal solutions. The performance of the HGA can be affected by the parameters of its operators. The optimal settings of these parameters will be obtained to find the best performance through a set of simulation on 36 benchmark functions with diverse properties. Through the evaluation on a set of benchmark functions, it was found that the HGA outperforms the MATLAB ga and particleswarm (PSO) functions in terms of the offline performance.