Neuro Fuzzy And Soft Computing Solution Manual

For all cases, the plans generated by the oFIS satisfied the given dose constraints (only the dose constraints and not the dose distributions were modified). The plans generated by the ANFIS were used as the starting point for the optimization in the subsequent manual and automated procedures and the plans generated by the oFIS were used as the starting point for the ANFIS. The most significant difference between the plans generated by the oFIS and the ANFIS was the number of required optimization iterations. Table 2 gives a summary of the number of optimization iterations (mean, standard deviation and range) for the plans generated by the different methods. The mean number of optimization iterations for the oFIS/ANFIS plans was significantly lower than for the oFIS plans. The mean optimizer performance time of the oFIS plans was significantly longer than for both other techniques. The difference in mean optimizer performance time between the oFIS and the ANFIS plans was insignificant.

For all cases, the plans generated by both oFIS and ANFIS fulfilled the given dose constraints (only the dose constraints and not the dose distributions were modified). The plans generated by the oFIS were used as the starting point for the optimization in the subsequent manual and automated procedures and the plans generated by the ANFIS were used as the starting point for the oFIS. The most significant difference between the plans generated by the oFIS and the ANFIS was the number of required optimization iterations. Table 1 gives a summary of the number of optimization iterations (mean, standard deviation and range), and optimizer performance times (mean, standard deviation and range) for the plans generated by the different methods. The mean number of optimization iterations for the oFIS/ANFIS plans was significantly lower than for the oFIS plans. The mean optimizer performance time of the oFIS plans was significantly longer than for both other techniques. The difference in mean optimizer performance time between the oFIS and the ANFIS plans was insignificant.

NFIDENT is able to create the input and output vectors for the ANFIS model. This provides the important capability to predict the output of the ANFIS model based on the input vectors for the ANFIS model, which is the basic function of the fuzzy inference system. This illustrates the potential of the hybrid learning approach, which combines a priori knowledge provided by the existing FIS with an automatic learning capability using the NEFPROX software. 827ec27edc