the analysis of concrete strength using rbf neural networks algorithm

performing a wide range of experiments in order to gain the most appropriate result to produce a product is both costly and time-consuming. therefore, using simulation and virtualization software for carrying out such experiments in frequent numbers and complex calculations is highly likely not only to save time and reduce costs, but also to increase the precision and accuracy of them. in the present study, the matlab software and neural networks have been applied to attain the percentage of microsilica for the highest strength of concrete. one of the most serious drawbacks of classifications using various neural networks is the presence of too many parameters to be taught. if such parameters are not appropriately opted, the efficiency is probably affected. one of the most frequently used ways to teach neural networks is trial and error to identify its parameters. in the present article, attempts have been made to optimize the number of parameters so that the data classification accuracy is increased using rbf neural network algorithm, which is considered to be one of the most popular artificial neural networks.