Development of adaptive real range (ARRange) genetic algorithms

We propose a new type of real genetic algorithms (GAs) named adaptive real range (ARRange) GAs. In conventional real GAs, we simply set a minimum and a maximum value for each design variable and divide the range into specific divisions after decoding binary strings to integers. However, in such cases, we need to give so many number of bits and divide the given range into a great number of discrete values in order to achieve sufficient accuracy. Moreover, initially, we usually do not have any information on the minimum and maximum values. Thus, we have to set them while ignoring the accuracy of the real values. In the proposed method, a range of real numbers will move adaptively in each generation by using the mean value and the standard deviation of the previous generation. In ARRange GAs, we do not have to consider the settings of minimum and maximum real values or number of bits for accuracy of real values. However, in ARRange GAs, we need four additional GA parameters that influence the performance of GAs. In particular, two of these parameters greatly influence the convergence. We also present additional options that relieves the designer from having to perform presettings. In this study, we demonstrate the proposed method by simple numerical examples and demonstrate its effectiveness and characteristics.