by Guest » Sun Jan 18, 2015 3:31 pm
Since [tex]a_{n+1}=F(a_n)[/tex], if the sequence is convergent with limit [tex]l[/tex], then this limit fulfills the equation [tex]l=F(l)[/tex]. The roots of this equation are 4 and 6.
First of all we will examine the values of [tex]a_n[/tex] for which [tex]a_{n+1}\ge a_n[/tex].
This means solving the inequality [tex]F(x)>x[/tex]. The idea is to use an important theorem for convergent sequences - a sequence which is monotone and bounded is convergent. The inequality holds for [tex]x \in (-12,4) \cup (6,+\infty)[/tex] respectively for [tex]a_n[/tex] in these intervals (*).
We will examine for what initial values [tex]a_1[/tex], [tex]a_2[/tex] lies in some of the 4 intervals (these intervals in which the real line is divided by the solution of the inequality).
1) [tex]a_2>6[/tex], so [tex]F(a_1)>6[/tex] and we find [tex]a_1 \in (-12,-4)\cup (6,+\infty)[/tex]. Then [tex]a_3=F(a_2)>a_2>6[/tex] and by induction [tex]a_{n+1}=F(a_n)>a_n>6[/tex] (we know that [tex]F(x)>x[/tex] for [tex]x>6[/tex]). Thus for these initial values the sequence is monotone. However it is not bounded - otherwise it would have limit which however is strictly greater then 6 - impossible as we saw in the beginning. Hence the sequence diverges.
2)[tex]6>a_2>4[/tex], so [tex]6>F(a_1)>4[/tex] and [tex]a_1\in (-4,-3)\cup (4,6)[/tex]. Thus [tex]a_n \in (4,6)[/tex] (since we see from the solutions that when [tex]x\in(4,6)[/tex] then [tex]F(x)\in(4,6)[/tex]). So for such initial value the sequence is bounded in [tex](4,6)[/tex] (after the first term). From (*) it follows that the sequence is decreasing. Hence its limit is 4.
3) [tex]4>a_2>-12[/tex], so [tex]4>F(a_1)>-12[/tex] and [tex]a_1\in (-3,4)[/tex]. In this case [tex]F(a_2)>a_2[/tex]. We should examine precisely where in [tex](-12,4)[/tex] [tex]a_2[/tex] lies in order to see where goes [tex]F(a_2)[/tex]. Observe that [tex]F(x)=a[/tex] has solution for [tex]a\in (-\inft,-94)\cup (2,+\infty)[/tex]. Thus [tex]a_2 \in (2,4)[/tex]. For such value we know that [tex]F(a_2)>a_2[/tex]. However [tex]F(a_2)<4[/tex] since [tex]a_2\ in (-3,4)[/tex](in the beginning of that point) ([tex]F(x)>x[/tex] for [tex]x\in (2,4)[/tex] and [tex]F(x)<4[/tex] in that interval). Inductively we proof that [tex]a_n<4[/tex] and is increasing, so the limit is 4.
4) [tex]a_2<-12[/tex], so [tex]F(a_1)<-12[/tex] and [tex]a_1\in (-\infty,-12)[/tex]. This means that [tex]a_n<-12[/tex] for all [tex]n[/tex] so the sequence cannot converge (the limit point candidates are 4 and 6). Actually it diverges.
Now it remains to write the answer
For [tex]a_1 \in (-12,-4)\cup (6,+\infty)[/tex] - divergent
For [tex]a_1\in (-4,-3)\cup (4,6)[/tex] - convergent with limit 4
For [tex]a_1\in (-3,4)[/tex] - convergent with limit 4
For [tex]a_1\in (-\infty,-12)[/tex] - divergent
For [tex]a_1=-3[/tex] and [tex]a_1=4[/tex] - convergent with limit 4
For [tex]a_1=-4[/tex] and [tex]a_1=6[/tex] - convergent with limit 6