Page 67 - Mathematical Techniques of Fractional Order Systems
P. 67

56  Mathematical Techniques of Fractional Order Systems



                                                 2
                                                     2:1 2
                                                           2
                           u 2 ðt; EÞ 52 0:455hð1:099t 1 ht E Þψ ;    ð2:77Þ
                                                              2 2:2 2
                             2
                                       2
                                                   2:1
                                                                      2 4:1 2
            u 3 ðt; EÞ 5 0:04134hψ ð2 12:0917t 2 22:0087ht E 2 9:9768h t E 1 h t ψ Þ;
                                                                      ð2:78Þ
            and so on. In this way can be obtained mth-order approximate solution for
            the problem (2.63) with initial condition (2.64) as follows
                                                 M
                                                X
                              uðtÞ  U M ðt; h; E; nÞ 5  u m ðt; EÞ:   ð2:79Þ
                                                m50
               According to the frame of the controlled Picard’s method the iterative
            formula (2.34), for Eqs. (2.63), (2.64) by choosing H(t) 5 1 is given by
                                 h  ð t   α21     α
                                                                       2
            u m11 ðt; h; EÞ; u m ðt; h; EÞ 2  ðt2τÞ  D ½u m ðτ; h; Eފðu m ðt; EÞÞ 2ðn11Þ  2 ψ dτ
                                               t
                               ΓðαÞ 0
                                                                       ð2:80Þ
               Starting with the initial approximation solution u 0 (t, E) 5 E which satisfies
            the initial condition (2.64), the successive approximations u m11 (t, h, E) for
            n 52 3is
                                                   1:9 2
                                u 1 ðt; h; EÞ 5 E 1 0:5472ht ψ ;      ð2:81Þ
                u 2 ðt; h; EÞ 5 1:0945ht ψ 2 0:5472h t E ψ 2 0:01292h t ψ 6
                                             2 1:9 2 2
                                                              4 5:7
                                1:9 2
                                                                      ð2:82Þ
                                        3 3:8 4
                         1 Eð1 2 0:11212h t ψ Þ;
            and for n 52 2, the first two solutions of Eq. (2.80) are given by
                                                    1:9 2
                               u 1 ðt; h; EÞ 5 E 1 0:5472 ht ψ ;      ð2:83Þ
                                                  4
                                                                  2 1:9 2
                                1:9 2
                                              3 3:8
               u 2 ðt; h; EÞ 5 1:0944ht ψ 2 0:05606h t ψ 1 Eð1 2 0:54724h t ψ Þ;
                                                                      ð2:84Þ
            and similarly for other iterations. With the help of the remaining boundary
            conditions (2.65), then the relation between the parameters h and E for the
            two methods is given by
                       uð1Þ  U M ð1; h; E; nÞ 5 1; uð1Þ  u m11 ð1; h; E; nÞ 5 1:  ð2:85Þ

               It has been shown in Magyari (2008) and Abbasbandy et al. (2009) that
            when n 52 2, the model (2.61) admits dual solutions for 0 , ψ #
            0.765152. To show that the number of solutions is continuous for the model
            in the fractional order domain the value of ψ variation about 0.7 is selected.
            Fig. 2.7A and B shows the h 2 E curve of Eq. (2.85), where two values of E
            are obtained which reflect the dual solutions in this case. It is clear from
            Fig. 2.7B the two line segment where E constant ended when ψ 5 0.74, this
   62   63   64   65   66   67   68   69   70   71   72