Selecting operators, selection strategy, and tuning parameters for genetic algorithms (GAs)is usually a very time-consuming job. In this article we introduce a method for developing an adaptive real-coded genetic algorithm (ARGA) which aims at reducing this computation time. In developing the algorithm, we first use factorial design experiments to identify ''important'' and ''sensitive'' parameters. Then these parameters will be dynamically changed during the evolutionary process by efficient computing budget allocation. At the end of the search process, not only has the optimum of the original problem been found, but also the adaptive changing pattern of the GA parameters has been captured. This algorithm is successfully used to solve some benchmark problems, a Linear-Quadratic-Gaussian (LQG) problem and a drug scheduling problem. The results show that ARGA outperforms simple GAs and other adaptive GAs which use the same type of operators. Moreover, ARGA is able to find the optimum for some difficult problems while the simple GAs with the best parameter combination can only reach the local optimum.