Extending Properties From Integer to Rational Exponents
While extending the notion of exponents from integers to rationals, we want to make sure of few things.
We want the laws of exponents for integer exponents to be valid for rational exponents.
We want to introduce notion of nth root or inverse function of f(x)=xn.
Lets try to define an1,n∈N. Consider (an)n1. As we want the laws of exponents to hold, we want (an)n1=an×n1=a1=a.
Do you have a fair idea what an1 means?
So, we want to define an1=na. In other words we want an1=b,b∈R where b is such that bn=a.
Lets try solving some equations of form bn=a.
Solve b5=3125.
Solve b2=169.
Do you see the problem while defining an1 as above?
There is a way to get away from this problem. We restrict a and also the value of ax,x∈Q to non negative real numbers. So now we want ∀a∈R≥0an1=b where b∈R≥0. But there is still a problem with defining an1 as we want to.
Do you see the problem in doing the above?
So now we define∀a∈R≥0an1=b where b∈R≥0 is such that bn=a.
As we want laws of exponents for integers to hold, we must define anm=(an1)m. These two definitions define the value of ax∀a∈R≥0and∀x∈Q.
Do you wish to go ahead?
Laws of exponents
Now we can prove the laws of exponents for rational exponents as stated below, proofs of which are quiet easy and follow from definitions and laws of integer exponents. Here a,b∈R≥0 and x,y∈Q
(ax)y=axy
a−x=ax1
ax×ay=ax+y
ax×bx=(a×b)x
If a<b and x>0 then ax<bx
Thus we now have managed to extend integer exponents to rational exponents so that the laws of exponents still hold and we have a way to express roots of real numbers.
But, as a price of this extension, we must restict the definition to only half of the rational numbers.
Time for some Practice...
4923=
8143=
1234×1831=
Existence and Uniqueness of an1
∀a∈R≥0∃!b∈R≥0 such that bn=a
Let X={y∈R≥0,yn≤a} Note that 0∈X and x is bounded above either by 1 or by a itself. Let b=sup(X). We claim that bn=a. Suppose not.
If bn<a then consider b+k1,k∈n such that k>a−bn2n−1ifb≤1 or k>a−bnbn−1(2n−1)ifb≥1.
If b≤1 then
(b+k1)n=bn+k1(1n)bn−1+k2(2n)bn−2+⋯+kn(nn)b0<bn+k2n−1<bn+a−bn=a
Thus b+k1∈X⇒b=sup(X)
If b≥1 then
(b+k1)n=bn+k1(1n)bn−1+k2(2n)bn−2+⋯+kn(nn)b0<bn+kbn−1(2n−1)<bn+a−bn=a Thus b+k1∈X⇒b=sup(X)
Similarly, if bn>a then we can take b−k1 where k is defined as above to guarantee that (b−k1)n>a⇒b=sup(x)
Suppose there is b such that bn=a then b=sup(X).
Hence, such b exists and is unique.