|
< ><FONT face="Times New Roman"><FONT size=3> <p></p></FONT></FONT></P>
< ><FONT face="Times New Roman" size=3>Series are a natural continuation of our study of functions. In the previous chapter we found how </FONT></P>
< ><FONT face="Times New Roman" size=3>to approximate our elementary functions by polynomials, with a certain error term. Conversely, one can define arbitrary functions by giving a series for them. We shall see how in the sections below.</FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> In practice, very few tests are used to determine convergence of series. Essentially, the comparision test is the most frequent. Furthermore, the most important series are those which converge absolutely. Thus we shall put greater emphasis on these.</FONT></FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> <p></p></FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>Convergent Series</FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> <p></p></FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>Suppose that we are given a sequcnce of numbers</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">a<SUB>1</SUB>,a<SUB>2</SUB>,a<SUB>3</SUB></FONT>…<FONT face="Times New Roman"> </FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">i.e. we are given a number a<SUB>n</SUB>, for each integer </FONT>n><FONT face="Times New Roman">1.We form the sums</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> S<SUB>n</SUB>=a<SUB>1</SUB>+a<SUB>2</SUB>+</FONT>…<FONT face="Times New Roman">+a<SUB>n<p></p></SUB></FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>It would be meaningless to form an infinite sum</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">a<SUB>1</SUB>+a<SUB>2</SUB>+a<SUB>3</SUB>+</FONT>…</FONT></P>
<P ><FONT face="Times New Roman" size=3>because we do not know how to add infinitely many numbers. However, if our sums S<SUB>n</SUB> approach a limit as n becomes large, then we say that the sum of our sequence converges, and we now define its sum to be that limit.</FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> The symbols </FONT></FONT></P>
<P ><FONT size=3>∑<SUB><FONT face="Times New Roman">a=1 </FONT></SUB><SUP>∞<FONT face="Times New Roman"> </FONT></SUP><FONT face="Times New Roman">a<SUB>n</SUB><SUP> <p></p></SUP></FONT></FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3><SUP> </SUP>will be called a series. We shall say that the series converges if the sums approach a limit as n becomes large. Otherwise, we say that it does not converge, or diverges. If the seriers converges, we say that the value of the series is </FONT></FONT></P>
<P ><FONT size=3>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">=lim<SUB>a</SUB></FONT><SUB>→∞</SUB><FONT face="Times New Roman">S<SUB>n</SUB>=lim<SUB>a</SUB></FONT><SUB>→∞</SUB><FONT face="Times New Roman">(a<SUB>1</SUB>+a<SUB>2</SUB>+</FONT>…<FONT face="Times New Roman">+a<SUB>n</SUB>)</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>In view of the fact that the limit of a sum is the sum of the limits, and other standard properties of limits, we get:</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">THEOREM 1. Let{ a<SUB>n</SUB> }and { b<SUB>n</SUB> }(n=1,2,</FONT>…<FONT face="Times New Roman">)</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>be two sequences and assume that the series </FONT></P>
<P ><FONT size=3>∑<FONT face="Times New Roman"><SUB>a=1</SUB><SUP> </SUP></FONT><SUP>∞<FONT face="Times New Roman"> </FONT></SUP><FONT face="Times New Roman">a<SUB>n</SUB></FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞<FONT face="Times New Roman"> </FONT></SUP><FONT face="Times New Roman">b<SUB>n</SUB><SUP> <p></p></SUP></FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">converge. Then </FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">(a<SUB>n </SUB>+ b<SUB>n </SUB>) also converges, and is equal to the sum of the two series. If c is a number, then </FONT></FONT></P>
<P ><FONT size=3>∑<FONT face="Times New Roman"> <SUB>a=1</SUB></FONT><SUP>∞</SUP><FONT face="Times New Roman">c a<SUB>n</SUB> =c</FONT>∑<FONT face="Times New Roman"><SUB>a=1</SUB> </FONT><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n<p></p></SUB></FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">Finally, if s<SUB>n</SUB>=a<SUB>1</SUB>+a<SUB>2</SUB>+</FONT>…<FONT face="Times New Roman">+a<SUB>n</SUB> and t<SUB>n</SUB>=b<SUB>1</SUB>+b<SUB>2</SUB>+</FONT>…<FONT face="Times New Roman">+b<SUB>n</SUB> then</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> </FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n </SUB></FONT>∑<SUB><FONT face="Times New Roman"> a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">b<SUB>n</SUB>=lim<SUB>a</SUB></FONT><SUB>→∞<FONT face="Times New Roman"> </FONT></SUB><FONT face="Times New Roman">s<SUB>n </SUB>t<SUB>n </SUB></FONT></FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> In particular, series can be added term by term. Of course , they cannot be multiplied term by term.</FONT></FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> We also observe that a similar theorem holds for the difference of two series.</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> If a series </FONT>∑<FONT face="Times New Roman">a<SUB>n</SUB> converges, then the numbers a<SUB>n</SUB> must approach 0 as n becomes large. However, there are examples of sequences {an} for which the series does not converge, and yet lim<SUB>a</SUB></FONT><SUB>→∞</SUB><FONT face="Times New Roman">a<SUB>n</SUB>=0</FONT></FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> <p></p></FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>Series with Positive Terms</FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> <p></p></FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">Throughout this section, we shall assume that our numbers a<SUB>n</SUB> are </FONT>><FONT face="Times New Roman"> 0. Then the partial sums</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> S<SUB>n</SUB>=a<SUB>1</SUB>+a<SUB>2</SUB>+</FONT>…<FONT face="Times New Roman">+a<SUB>n</SUB></FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>are increasing, i.e.</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">s<SUB>1</SUB></FONT><<FONT face="Times New Roman">s<SUB>2 </SUB></FONT><<FONT face="Times New Roman">s<SUB>3</SUB></FONT><…<<FONT face="Times New Roman">s<SUB>n</SUB></FONT><<FONT face="Times New Roman">s<SUB>n+1</SUB></FONT><…</FONT></P>
<P ><FONT face="Times New Roman" size=3>If they are approach a limit at all, they cannot become arbitrarily large. Thus in that case there is a number B such that </FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> S<SUB>n</SUB>< B</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>for all n. The collection of numbers {s<SUB>n</SUB>} has therefore a least upper bound ,i.e. there is a smallest number S such that </FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> s<SUB>n</SUB><S</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">for all n. In that case , the partial sums s<SUB>n</SUB> approach S as a limit. In other words, given any positive number </FONT>ε<FONT face="Times New Roman">>0, we have </FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">S –</FONT>ε<FONT face="Times New Roman">< s<SUB>n </SUB>< S</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>for all n .sufficiently large. This simply expresses the fact that S is the least of all upper bounds for our collection of numbers s<SUB>n</SUB>. We express this as a theorem.</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">THEOREM 2. Let{a<SUB>n</SUB>}(n=1,2,</FONT>…<FONT face="Times New Roman">)be a sequence of numbers>0 and let </FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> S<SUB>n</SUB>=a<SUB>1</SUB>+a<SUB>2</SUB>+</FONT>…<FONT face="Times New Roman">+a<SUB>n</SUB></FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>If the sequence of numbers {s<SUB>n</SUB>} is bounded, then it approaches a limit S , which is its least upper bound.</FONT></P>
<P ><FONT face="Times New Roman" size=3>Theorem 3 gives us a very useful criterion to determine when a series with positive terms converges:</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">THEOREM 3. Let</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n</SUB> and</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman"> b<SUB>n</SUB> be two series , with a<SUB>n</SUB>>0 for all n and b<SUB>n</SUB>>0 for all n. Assume that there is a number c such that </FONT></FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> a<SUB>n</SUB>< c b<SUB>n</SUB></FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">for all n, and that</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">b<SUB>n</SUB> converges. Then </FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman"> a<SUB>n</SUB> converges, and </FONT></FONT></P>
<P ><FONT size=3>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n </SUB>≤ c</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">b<SUB>n</SUB></FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>PROOF. We have </FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> a<SUB>1</SUB>+</FONT>…<FONT face="Times New Roman">+a<SUB>n</SUB></FONT>≤<FONT face="Times New Roman">cb<SUB>1</SUB>+</FONT>…<FONT face="Times New Roman">+cb<SUB>n</SUB></FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> =c(b<SUB>1</SUB>+</FONT>…<FONT face="Times New Roman">+b<SUB>n</SUB>)</FONT>≤<FONT face="Times New Roman"> c</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">b<SUB>n</SUB> </FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">This means that c</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">b<SUB>n</SUB> is a bound for the partial sums a<SUB>1</SUB>+</FONT>…<FONT face="Times New Roman">+a<SUB>n</SUB>.The least upper bound of these sums is therefore </FONT>≤<FONT face="Times New Roman"> c</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">b<SUB>n</SUB>, thereby proving our theorem.</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>Differentiation and Intergration of Power Series.</FONT></P>
<P ><FONT face="Times New Roman" size=3>If we have a polynomial</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> a<SUB>0</SUB>+a<SUB>1</SUB>x+</FONT>…<FONT face="Times New Roman">+a<SUB>n</SUB>x<SUP>n</SUP></FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">with numbers a<SUB>0</SUB>,a<SUB>1</SUB>,</FONT>…<FONT face="Times New Roman">,a<SUB>n</SUB> as coefficients, then we know how to find its derivative. It is a<SUB>1</SUB>+2a<SUB>2</SUB>x+</FONT>…<FONT face="Times New Roman">+na<SUB>n</SUB>x<SUP>n</SUP></FONT><SUP>–<FONT face="Times New Roman">1</FONT></SUP><FONT face="Times New Roman">. We would like to say that the derivative of a series can be taken in the same way, and that the derivative converges whenever the series does.</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> THEOREM 4. Let r be a number >0 and let </FONT>∑<FONT face="Times New Roman">a<SUB>n</SUB>x<SUP>n</SUP> be a series which converges absolutely for </FONT>∣<FONT face="Times New Roman">x</FONT>∣<FONT face="Times New Roman"><r. Then the series </FONT>∑<FONT face="Times New Roman">na<SUB>n</SUB>x<SUP>n-1</SUP> also converges absolutely for</FONT>∣<FONT face="Times New Roman">x</FONT>∣<FONT face="Times New Roman"><r.</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">A similar result holds for integration, but trivially. Indeed, if we have a series </FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n</SUB>x<SUP>n </SUP>which converges absolutely for </FONT>∣<FONT face="Times New Roman">x</FONT>∣<FONT face="Times New Roman"><r, then the series </FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> </FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n</SUB>/n+1 x<SUP>n+1</SUP>=x</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n</SUB>x<SUP>n</SUP><SUB> </SUB></FONT>∕<FONT face="Times New Roman">n+1 </FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>has terms whose absolute value is smaller than in the original series.</FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> The preceding result can be expressed by saying that an absolutely convergent series can be integrated and differentiated term by term and and still yields an absolutely convergent power series.</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>It is natural to expect that if </FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> f (x)=</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n</SUB>x<SUP>n</SUP>,</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>then f is differentiable and its derivative is given by differentiating the series term by term. The next theorem proves this.</FONT></P>
<P ><FONT face="Times New Roman"><FONT size=3> THEOREM 5. Let </FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> f (x)=</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman"> a<SUB>n</SUB>x<SUP>n</SUP></FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">be a power series, which converges absolutely for</FONT>∣<FONT face="Times New Roman">x</FONT>∣<FONT face="Times New Roman"><r. Then f is differentiable for </FONT>∣<FONT face="Times New Roman">x</FONT>∣<FONT face="Times New Roman"><r, and </FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> f</FONT>′<FONT face="Times New Roman">(x)=</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">na<SUB>n</SUB>x<SUP>n-1</SUP>.</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> THEOREM 6. Let f (x)=</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n</SUB>x<SUP>n</SUP> be a power series, which converges absolutely for </FONT>∣<FONT face="Times New Roman">x</FONT>∣<FONT face="Times New Roman"><r. Then the relation </FONT></FONT></P>
<P ><FONT size=3>∫<FONT face="Times New Roman">f (x)d x=</FONT>∑<SUB><FONT face="Times New Roman">a=1</FONT></SUB><SUP>∞</SUP><FONT face="Times New Roman">a<SUB>n</SUB>x<SUP>n+1</SUP></FONT>∕<FONT face="Times New Roman">n+1</FONT></FONT></P>
<P ><FONT size=3><FONT face="Times New Roman">is valid in the interval </FONT>∣<FONT face="Times New Roman">x</FONT>∣<FONT face="Times New Roman"><r.</FONT></FONT></P>
<P ><FONT face="Times New Roman" size=3>We omit the proofs of theorems 4,5 and 6.</FONT></P>
<P ><FONT size=3><FONT face="Times New Roman"> <p></p></FONT></FONT></P> |
|