SICP - Solution: Exercise 1.36

SICP - Solution: Exercise 1.36

October 28, 2018

Exercise 1.36 #

Modify fixed-point so that it prints the sequence of approximations it generates, using the newline and display primitives shown in Exercise 1.22. Then find a solution to ${x^x=1000}$ by finding a fixed point of $x\mapsto{\log(1000)/\log(x)}$. (Use Scheme’s primitive log procedure, which computes natural logarithms.) Compare the number of steps this takes with and without average damping. (Note that you cannot start fixed-point with a guess of 1, as this would cause division by ${\log(1)=0}$.)

Solution #

We can update fixed-point to show the sequence of approximation:

(define (fixed-point f first-guess)
  (define (close-enough? v1 v2)
    (< (abs (- v1 v2))
       tolerance))
  (define (try guess)
    (display guess) (newline)
    (let ((next (f guess)))
      (if (close-enough? guess next)
          next
          (try next))))
  (try first-guess))

Without damping:

(display (fixed-point (lambda (x) (/ (log 1000) (log x))) 2)) (newline)
2
9.965784284662087
3.004472209841214
6.279195757507157
3.759850702401539
5.215843784925895
4.182207192401397
4.8277650983445906
4.387593384662677
4.671250085763899
4.481403616895052
4.6053657460929
4.5230849678718865
4.577114682047341
4.541382480151454
4.564903245230833
4.549372679303342
4.559606491913287
4.552853875788271
4.557305529748263
4.554369064436181
4.556305311532999
4.555028263573554
4.555870396702851
4.555315001192079
4.5556812635433275
4.555439715736846
4.555599009998291
4.555493957531389
4.555563237292884
4.555517548417651
4.555547679306398
4.555527808516254
4.555540912917957
4.555532270803653

With damping, the evaluation becomes:

(display (fixed-point (lambda (x) (average x (/ (log 1000) (log x)))) 2)) (newline)
2
5.9828921423310435
4.922168721308343
4.628224318195455
4.568346513136242
4.5577305909237005
4.555909809045131
4.555599411610624
4.5555465521473675
4.555537551999825

It takes 35 steps to converge without damping, but only 10 steps with the damping method. It is clear that damping makes the convergence faster in this case.

Update #

Thanks to albondad for pointing out an error.