Math 5335     Section 2

Fall Semester 2005

Selected solutions: first HW assignment

 

Problem 18. The endpoints are   (4,3) - 5(-1,-2) = (9,13)   and   (4,3) - 2(-1,-2) = (6,7)   respectively. Therefore, the direction indicator for the segment is   (6,7) - (9,13) = (-3,-6),   and the segment is described as   {(9,13) + t(-3,-6): 0 t 1}.
 

Problem 26. The direction indicator of the ray is   (-4,-1) - (5,3) = (-9,-4).    The order of the subtraction is important. Finally, to get the unit direction indicator, as requested in the problem, we must multiply by the reciprocal of the length in order to get a vector of length 1. Thus, the answer is (-9,-4).
 

Problem 30. Let the line be given parametrically as   {P + sU : s },   and let the given points P0, P1, ..., Pn correspond to the parameter values s0, s1, ..., sn respectively.

We're given that Pj is between Pj-1 and Pj-1 for j = 1, ..., n-1. What this means for the parameter values is that for each   j   we have either   sj-1 < sj < sj+1   or   sj-1 > sj > sj+1. What we have to prove is that all of these inequalities line up coherently. More specifically, in the case where   s0 < s1 < s2   we have to prove that we have the following chain of inequalities:

s0 < s1 < s2 < ... < sn-1 < sn. From this chain of inequalities, together with transitivity of the "less than" relation for real numbers, we deduce that   si < sj < sk   whenever  i < j < k,   which implies the between-ness statement actually requested in the problem. (But normally, I wouldn't hold students in a course like this responsible for writing down this last particular bit of math terminology.)

{And strictly speaking, we also would have to prove that the opposite chain of inequalities holds in the case where   s0 > s1 > s2.   But that proof would be so similar to the first case that it can safely be omitted.}

So, we will prove by induction on   p,   starting with p = 2, that we have the following chain of inequalities:

s0 < s1 < s2 < ... < sp-1 < sp for every value of   p   where it makes sense. Then, the case p = n yeilds the main conclusion.

The initial step, namely p = 2, simply says that   s0 < s1 < s2.   This is true by assumption -- i.e., since we have decided to focus our discussion on the first case.

Thus, we proceed to the inductive step. Informally speaking, we have to show that if the chain of inequalities has been checked at the pth stage, then the corresponding chain at the (p+1)th stage will follow from this and the other given information. More formally, we assume that the chain of inequalities:

s0 < s1 < s2 < ... < sp-1 < sp holds, and we have to show that the [slightly longer] chain: s0 < s1 < s2 < ... < sp-1 < spsp+1 can be deduced from it and the other given information. The chain of inequalities at the pth stage is called the inductive hypothesis.

Well, taking  j = p  we have either:

Now, I claim that the second alternative contradicts our inductive hypothesis. Indeed, the second alternative includes the inequality   sp-1 > sp,   while our inductive hypothesis includes the equality   sp-1 < sp.  

Accordingly, the second alternative cannot hold. Therefore the first alternative holds. Since the first alternative includes the inequality   sp < sp+1,   we can combine it with the inductive hypothesis to obtain:

s0 < s1 < s2 < ... < sp-1 < spsp+1. This completes the inductive step, and thus also the proof.
 

Problem 40. Let's call the endpoints   P0 = (x1,x2)   and   P3 = (y1,y2).   One of the two trisection points is then:
      P1 = (x1,x2) + 1/3((y1,y2) - (x1,x2)) = . . . = 2/3(x1,x2) + 1/3(y1,y2).
{One or two steps were omitted from the calculation.}
If we do a similar calculation, starting with the coefficient   2/3   instead, then we find the other trisection point, namely:
      P2 = 1/3(x1,x2) + 2/3(y1,y2).
Now, these solutions look fairly "self-evident", but for completeness we should check that all 3 segments have the same length, namely that:
      ||Pi - Pi-1|| = 1/3((y1 - x1)² + (y2 - x2)1/2   for   i = 1, 2, 3.  
{The computations involved in doing this may be mildly tedious, but they are reasonably straightforward compared to some other things we've done lately.}


Back to the class homepage.


Prof. Joel Roberts
School of Mathematics
University of Minnesota
Minneapolis, MN 55455
USA

Office: 351 Vincent Hall
Phone: (612) 625-1076
Dept. FAX: (612) 626-2017
e-mail: roberts@math.umn.edu
http://www.math.umn.edu/~roberts