I have the following binary tree
A
/ \
B C
/ \
D E
represented as a list in Lisp (A 2 B 0 C 2 D 0 E 0) where the letters are node names and the numbers are the number of child nodes (0 for none, 1 one node, 2 two nodes). I need to find highest from root node to leaf depth of the tree (the depth of the binary tree that is) recursively. I'm pretty new to Lisp and I can't figure how to implement it. This is what I manage to come up with until now:
(defun depth (tree)
"Returns the depth of the argument tree."
(check-type tree list)
(if (= (second tree) 0)
0
(1+ (get-btree-max-depth (cddr tree)))))
(defun get-btree-max-depth (btree)
"Returns the maximum depth
of the argument tree."
(check-type btree list)
(if (= (second btree) 0)
0
(max (depth (cddr btree))
(get-btree-max-depth (cddr btree)))))
but it doesn't work properly. I also browsed similar postings but I didn't find anything useful that could help me. Could somebody give me a suggestion to help figure this out? Thank you!
P.S. This is part of a small project that I will present at University but also my own way of getting better in Lisp (I saw that many similar posts had questions asking if the posting is related to homework). :)
How about this one? No transformation of the tree needed.
(defun depth-rec (tree)
(labels ((depth-rec-aux (depth) ; self-recursive function
(if (null tree) ; no more nodes
depth ; -> return the current depth
(let ((n (second tree))) ; number of subnodes
(pop tree) (pop tree) ; remove the current node
(case n
(0 (1+ depth)) ; no subnode, 1+depth
(1 (depth-rec-aux (1+ depth))) ; one subnode, its depth+1
(2 (max (depth-rec-aux (1+ depth)) ; two subnodes, their max
(depth-rec-aux (1+ depth)))))))))
(depth-rec-aux 0))) ; start depth is 0
Another version:
(defun depth-rec (tree &aux (max 0))
(labels ((depth-rec-aux (depth)
(when tree
(pop tree)
(let ((n (pop tree)))
(if (zerop n)
(setf max (max max (1+ depth)))
(loop repeat n do (depth-rec-aux (1+ depth))))))))
(depth-rec-aux 0))
max)
I would first transform the list to a tree:
(defun tlist->tree (tlist)
"Transforms a tree represented as a kind of plist into a tree.
A tree like:
A
/ \
B C
/ / \
F D E
would have a tlist representation of (A 2 B 1 F 0 C 2 D 0 E 0).
The tree representation would be (A (B (F)) (C (D) (E)))"
(let (tree)
(push (pop tlist) tree)
(dotimes (i (pop tlist))
(multiple-value-bind (subnode rest-tlist) (tlist->tree tlist)
(push subnode tree)
(setf tlist rest-tlist)))
(values (nreverse tree) tlist)))
I wonder if you couldn't start with this tree representation to begin with.
Then, finding the depth of a tree in tree representation is a simple recursive one-liner.
Here's one in continuation-passing style:
(defun oddtree-height (oddtree)
(suboddtree-height oddtree
#'(lambda (h remainder)
(if (null remainder) h nil))))
(defun suboddtree-height (oddtree c)
(max-height-of-suboddtrees (cadr oddtree)
0
(cddr oddtree)
#'(lambda (h remainder)
(funcall c (+ h 1) remainder))))
(defun max-height-of-suboddtrees (n best oddtree c)
(if (= n 0)
(funcall c best oddtree)
(suboddtree-height oddtree
#'(lambda (h remainder)
(max-height-of-suboddtrees (- n 1) (max best h) remainder c)))))
Using Artelius's and Svante's answer I managed to solve the issue. Here is the code, perhaps it will be of some help to somebody else in need.
(defun btree-max-depth (btree)
"Returns the maximum depth
of the binary tree."
(check-type btree list)
(if (null btree)
0 ; the max depth of the members of ()
(max (depth (first btree))
(btree-max-depth (rest btree)))))
(defun depth (tree)
"Returns the depth of the argument TREE."
(if (atom tree)
0 ; an atomic tree has a depth of 0
(1+ (btree-max-depth tree))))
Thanks Artelius and Svante for your help!
Related
I'm arduously struggling my way through the N-queens problem in SICP (the book; I spent a few days on it -- last question here: Solving Eight-queens in scheme). Here is what I have for the helper functions:
#lang sicp
; the SICP language in Racket already defines this:
; (define nil '()
; boilerplate: filter function and range functions
(define (filter func lst)
(cond
((null? lst)
nil)
(else
(if (func (car lst))
(cons (car lst) (filter func (cdr lst)))
(filter func (cdr lst))))))
(define (range a b)
(if (> a b)
nil
(cons a (range (+ 1 a) b))))
; Selectors/handlers to avoid confusion on the (col, row) notation:
; representing it a position as (col, row), using 1-based indexing
(define (make-position col row) (cons col (list row)))
(define (col p) (car p))
(define (row p) (cadr p))
; adding a new position to a board
(define (add-new-position existing-positions p)
(append existing-positions
(list (make-position (col p) (row p)))))
; The 'safe' function
(define (any? l proc)
(cond ((null? l) #f)
((proc (car l)) #t)
(else (any? (cdr l) proc))))
(define (none? l proc) (not (any? l proc)))
(define (safe? existing-positions p)
(let ((bool (lambda (x) x)) (r (row p)) (c (col p)))
(and
; is the row safe? i.e., no other queen occupies that row?
(none? (map (lambda (p) (= (row p) r)) existing-positions)
bool)
; safe from the diagonal going up
(none? (map (lambda (p) (= r (+ (row p) (- c (col p)))))
existing-positions)
bool)
; safe from the diagonal going down
(none? (map (lambda (p) (= r (- (row p) (- c (col p)))))
existing-positions)
bool))))
And now, with that boilerplate, the actual/monstrous first working version I have of the queens problem:
(define (positions-for-col col size)
(map (lambda (ri) (make-position col ri))
(range 1 size)))
(define (queens board-size)
(define possible-positions '())
(define safe-positions '())
(define all-new-position-lists '())
(define all-positions-list '())
; existing-positions is a LIST of pairs
(define (queen-cols col existing-positions)
(if (> col board-size)
(begin
(set! all-positions-list
(append all-positions-list (list existing-positions))))
(begin
; for the column, generate all possible positions,
; for example (3 1) (3 2) (3 3) ...
(set! possible-positions (positions-for-col col board-size))
; (display "Possible positions: ") (display possible-positions) (newline)
; filter out the positions that are not safe from existing queens
(set! safe-positions
(filter (lambda (pos) (safe? existing-positions pos))
possible-positions))
; (display "Safe positions: ") (display safe-positions) (newline)
(if (null? safe-positions)
; bail if we don't have any safe positions
'()
; otherwise, build a list of positions for each safe possibility
; and recursively call the function for the next column
(begin
(set! all-new-position-lists
(map (lambda (pos)
(add-new-position existing-positions pos))
safe-positions))
; (display "All positions lists: ") (display all-new-position-lists) (newline)
; call itself for the next column
(map (lambda (positions-list) (queen-cols (+ 1 col)
positions-list))
all-new-position-lists))))))
(queen-cols 1 '())
all-positions-list)
(queens 5)
(((1 1) (2 3) (3 5) (4 2) (5 4))
((1 1) (2 4) (3 2) (4 5) (5 3))
((1 2) (2 4) (3 1) (4 3) (5 5))
((1 2) (2 5) (3 3) (4 1) (5 4))
((1 3) (2 1) (3 4) (4 2) (5 5))
To be honest, I think I did all the set!s so that I could more easily debug things (is that common?) How could I remove the various set!s to make this a proper functional-procedure?
As an update, the most 'terse' I was able to get it is as follows, though it still appends to a list to build the positions:
(define (queens board-size)
(define all-positions-list '())
(define (queen-cols col existing-positions)
(if (> col board-size)
(begin
(set! all-positions-list
(append all-positions-list
(list existing-positions))))
(map (lambda (positions-list)
(queen-cols (+ 1 col) positions-list))
(map (lambda (pos)
(add-new-position existing-positions pos))
(filter (lambda (pos)
(safe? existing-positions pos))
(positions-for-col col board-size))))))
(queen-cols 1 nil)
all-positions-list)
Finally, I think here is the best I can do, making utilization of a 'flatmap' function that helps deal with nested lists:
; flatmap to help with reduction
(define (reduce function sequence initializer)
(let ((elem (if (null? sequence) nil (car sequence)))
(rest (if (null? sequence) nil (cdr sequence))))
(if (null? sequence)
initializer
(function elem
(reduce function rest initializer)))))
(define (flatmap proc seq)
(reduce append (map proc seq) nil))
; actual
(define (queens board-size)
(define (queen-cols col existing-positions)
(if (> col board-size)
(list existing-positions)
(flatmap
(lambda (positions-list)
(queen-cols (+ 1 col) positions-list))
(map
(lambda (pos)
(add-new-position existing-positions
pos))
(filter
(lambda (pos)
(safe? existing-positions pos))
(positions-for-col col board-size))))))
(queen-cols 1 nil))
Are there any advantages of this function over the one using set! or is it more a matter of preference (I find the set! one easier to read and debug).
When you are doing the SICP problems, it would be most beneficial if you strive to adhere to the spirit of the question. You can determine the spirit from the context: the topics covered till the point you are in the book, any helper code given, the terminology used etc. Specifically, avoid using parts of the scheme language that have not yet been introduced; the focus is not on whether you can solve the problem, it is on how you solve it. If you have been provided helper code, try to use it to the extent you can.
SICP has a way of building complexity; it does not introduce a concept unless it has presented enough motivation and justification for it. The underlying theme of the book is simplification through abstraction, and in this particular section you are introduced to various higher order procedures -- abstractions like accumulate, map, filter, flatmap which operate on sequences/lists, to make your code more structured, compact and ultimately easier to reason about.
As illustrated in the opening of this section, you could very well avoid the use of such higher programming constructs and still have programs that run fine, but their (liberal) use results in more structured, readable, top-down style code. It draws parallels from the design of signal processing systems, and shows how we can take inspiration from it to add structure to our code: using procedures like map, filter etc. compartmentalize our code's logic, not only making it look more hygienic but also more comprehensible.
If you prematurely use techniques which don't come until later in the book, you will be missing out on many key learnings which the authors intend for you from the present section. You need to shed the urge to think in an imperative way. Using set! is not a good way to do things in scheme, until it is. SICP forces you down a 'difficult' path by making you think in a functional manner for a reason -- it is for making your thinking (and code) elegant and 'clean'.
Just imagine how much more difficult it would be to reason about code which generates a tree recursive process, wherein each (child) function call is mutating the parameters of the function. Also, as I mentioned in the comments, assignment places additional burden upon the programmers (and on those who read their code) by making the order of the expressions have a bearing on the results of the computation, so it is harder to verify that the code does what is intended.
Edit: I just wanted to add a couple of points which I feel would add a bit more insight:
Your code using set! is not wrong (or even very inelegant), it is just that in doing so, you are being very explicit in telling what you are doing. Iteration also reduces the elegance a bit in addition to being bottom up -- it is generally harder to think bottom up.
I feel that teaching to do things recursively where possible is one of the aims of the book. You will find that recursion is a crucial technique, the use of which is inevitable throughout the book. For instance, in chapter 4, you will be writing evaluators (interpreters) where the authors evaluate the expressions recursively. Even much earlier, in section 2.3, there is the symbolic differentiation problem which is also an exercise in recursive evaluation of expressions. So even though you solved the problem imperatively (using set!, begin) and bottom-up iteration the first time, it is not the right way, as far as the problem statement is concerned.
Having said all this, here is my code for this problem (for all the structure and readability imparted by FP, comments are still indispensable):
; the board is a list of lists - a physical n x n board, where
; empty positions are 0 and filled positions are 1
(define (queens board-size)
(let ((empty-board (empty-board-gen board-size))) ; minor modification - making empty-board available to queen-cols
(define (queen-cols k)
(if (= k 0)
(list empty-board)
(filter (lambda (positions) (safe? k positions))
; the flatmap below generates a list of new positions
; by 'adjoining position'- adding 'board-size' number
; of new positions for each of the positions obtained
; recursively from (queen-cols (- k 1)), which have
; been found to be safe till column k-1. This new
; set (list) of positions is then filtered using the
; safe? function to filter out unsafe positions
(flatmap
(lambda (rest-of-queens)
; the map below adds 'board-size' number of new
; positions to 'rest-of-queens', which is an
; element of (queen-cols (- k 1))
(map (lambda (new-row)
(adjoin-position new-row k rest-of-queens))
(enumerate-interval 1 board-size)))
(queen-cols (- k 1))))))
(queen-cols board-size)) ; end of let block
)
; add a column having a queen placed at position (new-row, col).
(define (adjoin-position new-row col rest-queens)
(let ((board-dim (length rest-queens))) ;length of board
; first create a zero 'vector', put a queen in it at position
; 'new-row', then put (replace) this new vector/column at the
; 'col' position in rest-queens
(replace-elem (replace-elem 1 new-row (gen-zero-vector board-dim)) col rest-queens)))
(define (safe? k positions) ; the safe function
(let ((row-pos-k (non-zero-index (item-at-index k positions)))) ; get the row of the queen in column k
(define (iter-check col rem) ;iteratively check if column 'col' of the board is safe wrt the kth column
(let ((rw-col (non-zero-index (car rem)))) ; get the row of 'col' in which a queen is placed
(cond ((= k 1) #t); 1x1 board is always safe
((= col k) #t); if we reached the kth column, we are done
; some simple coordinate geometry
; checks if the row of the queen in col and kth
; column is same, and also checks if the 'slope' of
; the line connecting the queens of the two columns
; is 1 (i.e. if it's a diagonal), if either is true,
; the kth queen is not safe
((or (= row-pos-k rw-col) (= (- k col) (abs (- row-pos-k rw-col)))) #f)
(else (iter-check (+ col 1) (cdr rem)))))) ; check the next column
(iter-check 1 positions))) ; start checking from the first column
; helper functions follow
(define (item-at-index n items) ; given a list, return the nth element
(define (iter idx rem)
(if (= idx n)
(car rem)
(iter (+ idx 1) (cdr rem))))
(iter 1 items))
(define (non-zero-index items)
; gives the first non-zero element from items - used for
; determining the row at which a queen is placed
(define (iter a rem)
(if (> (car rem) 0)
a
(iter (+ a 1) (cdr rem))))
(iter 1 items))
(define (empty-board-gen n) ; the empty board is n lists, each list with n zeros
(map (lambda (x) (map (lambda (y) 0) (enumerate-interval 1 n))) (enumerate-interval 1 n)))
(define (replace-elem new-elem pos items) ; replace item at position pos in items by new-elem, ultimately used for replacing an empty column with a column which has a queen
(define (iter i res rem)
(if (= i pos)
(append res (list new-elem) (cdr rem))
(iter (+ i 1) (append res (list(car rem))) (cdr rem)))) (iter 1 '() items))
(define (gen-zero-vector n) ; generate a list of length n with only zeros as elements
(define (iter a res)
(if (> a n)
res
(iter (+ a 1) (append res (list 0))))) (iter 1 '()))
(define (flatmap proc seq)
(accumulate append '() (map proc seq)))
(define (length items) ; not particularly efficient way for length of a list
(accumulate + 0 (map (lambda (x) 1) items)))
(define (accumulate op null-value seq)
(if (null? seq)
null-value
(op (car seq) (accumulate op null-value (cdr seq)))))
(define (enumerate-interval low high) ; a list of integers from low to hi
(define (iter a b res)
(if (> a b)
res
(iter (+ a 1) b (append res (cons a '())))))
(iter low high '()))
There are many ways to tackle this problem. I'll attempt to write a short and concise solution using Racket-specific procedures, explaining each step of the way. A solution using only the Scheme procedures explained in SICP is also possible, but it'll be more verbose and I'd argue, more difficult to understand.
My aim is to write a functional-programming style solution reusing as many built-in procedures as possible, and avoiding mutation at all costs - this is the style that SICP encourages you to learn. I'll deviate from the template solution in SICP if I think we can get a clearer solution by reusing existing Racket procedures (it follows then, that this code must be executed using the #lang racket language), but I've provided another answer that fits exactly exercise 2.42 in the book, implemented in standard Scheme and compatible with #lang sicp.
First things first. Let's agree on how are we going to represent the board - this is a key point, the way we represent our data will have a big influence on how easy (or hard) is to implement our solution. I'll use a simple representation, with only the minimum necessary information.
Let's say a "board" is a list of row indexes. My origin of coordinates is the position (0, 0), on the top-left corner of the board. For the purpose of this exercise we only need to keep track of the row a queen is in, the column is implicitly represented by its index in the list and there can only be one queen per column. Using my representation, the list '(2 0 3 1) encodes the following board, notice how the queens' position is uniquely represented by its row number and its index:
0 1 2 3
0 . Q . .
1 . . . Q
2 Q . . .
3 . . Q .
Next, let's see how are we going to check if a new queen added at the end of the board is "safe" with respect to the previously existing queens. For this, we need to check if there are any other queens in the same row, or if there are queens in the diagonal lines starting from the new queen's position. We don't need to check for queens in the same column, we're trying to set a single new queen and there aren't any others in this row. Let's split this task in multiple procedures.
; main procedure for checking if a queen in the given
; column is "safe" in the board; there are no more
; queens to the "right" or in the same column
(define (safe? col board)
; we're only interested in the queen's row for the given column
(let ([row (list-ref board (sub1 col))])
; the queen must be safe on the row and on the diagonals
(and (safe-row? row board)
(safe-diagonals? row board))))
; check if there are any other queens in the same row,
; do this by counting how many times `row` appears in `board`
(define (safe-row? row board)
; only the queen we want to add can be in this row
; `curry` is a shorthand for writing a lambda that
; compares `row` to each element in `board`
(= (count (curry equal? row) board) 1))
; check if there are any other queens in either the "upper"
; or the "lower" diagonals starting from the current queen's
; position and going to the "left" of it
(define (safe-diagonals? row board)
; we want to traverse the row list from right-to-left so we
; reverse it, and remove the current queen from it; upper and
; lower positions are calculated starting from the current queen
(let loop ([lst (rest (reverse board))]
[upper (sub1 row)]
[lower (add1 row)])
; the queen is safe after checking all the list
(or (null? lst)
; the queen is not safe if we find another queen in
; the same row, either on the upper or lower diagonal
(and (not (= (first lst) upper))
(not (= (first lst) lower))
; check the next position, updating upper and lower
(loop (rest lst) (sub1 upper) (add1 lower))))))
Some optimizations could be done, for example stopping early if there's more than one queen in the same row or stopping when the diagonals' rows fall outside of the board, but they'll make the code harder to understand and I'll leave them as an exercise for the reader.
In the book they suggest we use an adjoin-position procedure that receives both row and column parameters; with my representation we only need the row so I'm renaming it to add-queen, it simply adds a new queen at the end of a board:
; add a new queen's row to the end of the board
(define (add-queen queen-row board)
(append board (list queen-row)))
Now for the fun part. With all of the above procedures in place, we need to try out different combinations of queens and filter out those that are not safe. We'll use higher-order procedures and recursion for implementing this backtracking solution, there's no need to use set! at all as long as we're in the right mindset.
This will be easier to understand if you read if from the "inside out", try to grok what the inner parts do before going to the outer parts, and always remember that we're unwinding our way in a recursive process: the first case that will get executed is when we have an empty board, the next case is when we have a board with only one queen in position and so on, until we finally have a full board.
; main procedure: returns a list of all safe boards of the given
; size using our previously defined board representation
(define (queens board-size)
; we need two values to perform our computation:
; `queen-col`: current row of the queen we're attempting to set
; `board-size`: the full size of the board we're trying to fill
; I implemented this with a named let instead of the book's
; `queen-cols` nested procedure
(let loop ([queen-col board-size])
; if there are no more columns to try exit the recursion
(if (zero? queen-col)
; base case: return a list with an empty list as its only
; element; remember that the output is a list of lists
; the book's `empty-board` is just the empty list '()
(list '())
; we'll generate queen combinations below, but only the
; safe ones will survive for the next recursive call
(filter (λ (board) (safe? queen-col board))
; append-map will flatten the results as we go, we want
; a list of lists, not a list of lists of lists of...
; this is equivalent to the book's flatmap implementation
(append-map
(λ (previous-boards)
(map (λ (new-queen-row)
; add a new queen row to each one of
; the previous valid boards we found
(add-queen new-queen-row previous-boards))
; generate all possible queen row values for this
; board size, this is similar to the book's
; `enumerate-interval` but starting from zero
(range board-size)))
; advance the recursion, try a smaller column
; position, as the recursion unwinds this will
; return only previous valid boards
(loop (sub1 queen-col)))))))
And that's all there is to it! I'll provide a couple of printing procedures (useful for testing) which should be self-explanatory; they take my compact board representation and print it in a more readable way. Queens are represented by 'o and empty spaces by 'x:
(define (print-board board)
(for-each (λ (row) (printf "~a~n" row))
(map (λ (row)
(map (λ (col) (if (= row col) 'o 'x))
board))
(range (length board)))))
(define (print-all-boards boards)
(for-each (λ (board) (print-board board) (newline))
boards))
We can verify that things work and that the number of solutions for the 8-queens problem is as expected:
(length (queens 8))
=> 92
(print-all-boards (queens 4))
(x x o x)
(o x x x)
(x x x o)
(x o x x)
(x o x x)
(x x x o)
(o x x x)
(x x o x)
As a bonus, here's another solution that works with the exact definition of queens as provided in the SICP book. I won't go into details because it uses the same board representation (except that here the indexes start in 1 not in 0) and safe? implementation of my previous answer, and the explanation for the queens procedure is essentially the same. I did some minor changes to favor standard Scheme procedures, so hopefully it'll be more portable.
#lang racket
; redefine procedures already explained in the book with
; Racket equivalents, delete them and use your own
; implementation to be able to run this under #lang sicp
(define flatmap append-map)
(define (enumerate-interval start end)
(range start (+ end 1)))
; new definitions required for this exercise
(define empty-board '())
(define (adjoin-position row col board)
; `col` is unused
(append board (list row)))
; same `safe?` implementation as before
(define (safe? col board)
(let ((row (list-ref board (- col 1))))
(and (safe-row? row board)
(safe-diagonals? row board))))
(define (safe-row? row board)
; reimplemented to use standard Scheme procedures
(= (length (filter (lambda (r) (equal? r row)) board)) 1))
(define (safe-diagonals? row board)
(let loop ((lst (cdr (reverse board)))
(upper (- row 1))
(lower (+ row 1)))
(or (null? lst)
(and (not (= (car lst) upper))
(not (= (car lst) lower))
(loop (cdr lst) (- upper 1) (+ lower 1))))))
; exact same implementation of `queens` as in the book
(define (queens board-size)
(define (queen-cols k)
(if (= k 0)
(list empty-board)
(filter
(lambda (positions) (safe? k positions))
(flatmap
(lambda (rest-of-queens)
(map (lambda (new-row)
(adjoin-position new-row k rest-of-queens))
(enumerate-interval 1 board-size)))
(queen-cols (- k 1))))))
(queen-cols board-size))
; debugging
(define (print-board board)
(for-each (lambda (row) (display row) (newline))
(map (lambda (row)
(map (lambda (col) (if (= row col) 'o 'x))
board))
(enumerate-interval 1 (length board)))))
(define (print-all-boards boards)
(for-each (lambda (board) (print-board board) (newline))
boards))
The above code is more in spirit with the original exercise, which asked you to implement just three definitions: empty-board, adjoin-position and safe?, thus this was more of a question about data representation. Unsurprisingly, the results are the same:
(length (queens 8))
=> 92
(print-all-boards (queens 4))
(x x o x)
(o x x x)
(x x x o)
(x o x x)
(x o x x)
(x x x o)
(o x x x)
(x x o x)
I want to write a function in Racket which takes an amount of money and a list of specific bill-values, and then returns a list with the amount of bills used of every type to make the given amount in total. For example (calc 415 (list 100 10 5 2 1)) should return '(4 1 1 0 0).
I tried it this way but this doesn't work :/ I think I haven't fully understood what you can / can't do with set! in Racket, to be honest.
(define (calc n xs)
(cond ((null? xs) (list))
((not (pair? xs))
(define y n)
(begin (set! n (- n (* xs (floor (/ n xs)))))
(list (floor (/ y xs))) ))
(else (append (calc n (car xs))
(calc n (cdr xs))))))
Your procedure does too much and you use mutation which is uneccesary. If you split the problem up.
(define (calc-one-bill n bill)
...)
;; test
(calc-one-bill 450 100) ; ==> 4
(calc-one-bill 450 50) ; ==> 9
Then you can make:
(define (calc-new-n n bill amount)
...)
(calc-new-n 450 100 4) ; ==> 50
(calc-new-n 450 50 9) ; ==> 0
Then you can reduce your original implememntation like this:
(define (calc n bills)
(if (null? bills)
(if (zero? n)
'()
(error "The unit needs to be the last element in the bills list"))
(let* ((bill (car bills))
(amount (calc-one-bill n bill)))
(cons amount
(calc (calc-new-n n bill amount)
(cdr bills))))))
This will always choose the solution with fewest bills, just as your version seems to do. Both versions requires that the last element in the bill passed is the unit 1. For a more complex method, that works with (calc 406 (list 100 10 5 2)) and that potentially can find all combinations of solutions, see Will's answer.
This problem calls for some straightforward recursive non-deterministic programming.
We start with a given amount, and a given list of bill denominations, with unlimited amounts of each bill, apparently (otherwise, it'd be a different problem).
At each point in time, we can either use the biggest bill, or not.
If we use it, the total sum lessens by the bill's value.
If the total is 0, we've got our solution!
If the total is negative, it is invalid, so we should abandon this path.
The code here will follow another answer of mine, which finds out the total amount of solutions (which are more than one, for your example as well). We will just have to mind the solutions themselves as well, whereas the code mentioned above only counted them.
We can code this one as a recursive-backtracking procedure, calling a callback with each successfully found solution from inside the deepest level of recursion (tantamount to the most deeply nested loop in the nested loops structure created with recursion, which is the essence of recursive backtracking):
(define (change sum bills callback)
(let loop ([sum sum] [sol '()] [bills bills]) ; "sol" for "solution"
(cond
((zero? sum) (callback sol)) ; process a solution found
((< sum 0) #f)
((null? bills) #f)
(else
(apply
(lambda (b . bs) ; the "loop":
;; 1. ; either use the first
(loop (- sum b) (cons b sol) bills) ; denomination,
;; 2. ; or,
(loop sum sol bs)) ; after backtracking, don't!
bills)))))
It is to be called through e.g. one of
;; construct `the-callback` for `solve` and call
;; (solve ...params the-callback)
;; where `the-callback` is an exit continuation
(define (first-solution solve . params)
(call/cc (lambda (return)
(apply solve (append params ; use `return` as
(list return)))))) ; the callback
(define (n-solutions n solve . params) ; n assumed an integer
(let ([res '()]) ; n <= 0 gets ALL solutions
(call/cc (lambda (break)
(apply solve (append params
(list (lambda (sol)
(set! res (cons sol res))
(set! n (- n 1))
(cond ((zero? n) (break)))))))))
(reverse res)))
Testing,
> (first-solution change 406 (list 100 10 5 2))
'(2 2 2 100 100 100 100)
> (n-solutions 7 change 415 (list 100 10 5 2 1))
'((5 10 100 100 100 100)
(1 2 2 10 100 100 100 100)
(1 1 1 2 10 100 100 100 100)
(1 1 1 1 1 10 100 100 100 100)
(5 5 5 100 100 100 100)
(1 2 2 5 5 100 100 100 100)
(1 1 1 2 5 5 100 100 100 100))
Regarding how this code is structured, cf. How to generate all the permutations of elements in a list one at a time in Lisp? It creates nested loops with the solution being accessible in the innermost loop's body.
Regarding how to code up a non-deterministic algorithm (making all possible choices at once) in a proper functional way, see How to do a powerset in DrRacket? and How to find partitions of a list in Scheme.
I solved it this way now :)
(define (calc n xs)
(define (calcAssist n xs usedBills)
(cond ((null? xs) usedBills)
((pair? xs)
(calcAssist (- n (* (car xs) (floor (/ n (car xs)))))
(cdr xs)
(append usedBills
(list (floor (/ n (car xs)))))))
(else
(if ((= (- n (* xs (floor (/ n xs)))) 0))
(append usedBills (list (floor (/ n xs))))
(display "No solution")))))
(calcAssist n xs (list)))
Testing:
> (calc 415 (list 100 10 5 2 1))
'(4 1 1 0 0)
I think this is the first program I wrote when learning FORTRAN! Here is a version which makes no bones about using everything Racket has to offer (or, at least, everything I know about). As such it's probably a terrible homework solution, and it's certainly prettier than the FORTRAN I wrote in 1984.
Note that this version doesn't search, so it will get remainders even when it does not need to. It never gets a remainder if the lowest denomination is 1, of course.
(define/contract (denominations-of amount denominations)
;; split amount into units of denominations, returning the split
;; in descending order of denomination, and any remainder (if there is
;; no 1 denomination there will generally be a remainder).
(-> natural-number/c (listof (integer-in 1 #f))
(values (listof natural-number/c) natural-number/c))
(let handle-one-denomination ([current amount]
[remaining-denominations (sort denominations >)]
[so-far '()])
;; handle a single denomination: current is the balance,
;; remaining-denominations is the denominations left (descending order)
;; so-far is the list of amounts of each denomination we've accumulated
;; so far, which is in ascending order of denomination
(if (null? remaining-denominations)
;; we are done: return the reversed accumulator and anything left over
(values (reverse so-far) current)
(match-let ([(cons first-denomination rest-of-the-denominations)
remaining-denominations])
(if (> first-denomination current)
;; if the first denomination is more than the balance, just
;; accumulate a 0 for it and loop on the rest
(handle-one-denomination current rest-of-the-denominations
(cons 0 so-far))
;; otherwise work out how much of it we need and how much is left
(let-values ([(q r)
(quotient/remainder current first-denomination)])
;; and loop on the remainder accumulating the number of bills
;; we needed
(handle-one-denomination r rest-of-the-denominations
(cons q so-far))))))))
I'm new to Lisp and I'm trying to solve an 8-puzzle using simple dfs (depth-first search).
But I am getting a program stack overflow.
My code:
(setq used (list))
(defun is_used (state lst)
(cond
((null lst) nil)
((equalp (car lst) state) t)
(t (is_used state (cdr lst)))))
(defun move (lst direction)
(let* ( (zero (find_zero lst))
(row (floor zero 3))
(col (mod zero 3))
(res (copy-list lst)))
(cond
((eq direction 'L)
(if (> col 0)
(rotatef (elt res zero) (elt res (- zero 1)))))
((eq direction 'R)
(if (< col 2)
(rotatef (elt res zero) (elt res (+ zero 1)))))
((eq direction 'U)
(if (> row 0)
(rotatef (elt res zero) (elt res (- zero 3)))))
((eq direction 'D)
(if (< row 2)
(rotatef (elt res zero) (elt res (+ zero 3))))))
(if (equalp res lst)
(return-from move nil))
(return-from move res))
nil)
(defun dfs (cur d prev)
; (write (length used))
; (terpri)
(push cur used)
(let* ((ways '(L R U D)))
(loop for dir in ways
do (if (move cur dir)
(if (not (is_used (move cur dir) used))
(dfs (move cur dir) (+ d 1) cur))))))
state here is a list of 9 atoms.
With uncommented (write (length used)) it prints 723 - number of items in used before the stack overflow occurs.
Now, before solving 8-puzzle, I just want to iterate over all possible states (there are exactly 9! / 2 = 181440 possible states). Sure, it may take some time, but how can I avoid the stack overflow here?
This is a typical problem explained in some AI programming books. If you need to search a large / unbounded amount of states, you should not use recursion. Recursion in CL is limited by the stack depth. Some implementations can optimize tail recursion - but then you need architecture your code, so that it is tail recursive.
Typically a data structure for that will be called an agenda. It keeps the states still to explore. If you look at a state, you push all states to explore from there onto the agenda. Make sure the agenda is in some way sorted (this might determine if it is depths or breadths first). Then take the next state from the agenda and examine it. If the goal is reached, then you are done. If the agenda is empty before the goal is found, then there is no solution. Otherwise loop...
Your code, simplified, is
(setq *used* (list))
(defun move (position direction)
(let* ( (zero (position 0 position))
(row (floor zero 3))
(col (mod zero 3))
(command (find direction `((L ,(> col 0) ,(- zero 1))
(R ,(< col 2) ,(+ zero 1))
(U ,(> row 0) ,(- zero 3))
(D ,(< row 2) ,(+ zero 3)))
:key #'car)))
(if (cadr command)
(let ((res (copy-list position)))
(rotatef (elt res zero) (elt res (caddr command)))
res))))
(defun dfs-rec (cur_pos depth prev_pos)
(write (length *used*)) (write '_) (write depth) (write '--)
; (terpri)
(push cur_pos *used*)
(let* ((dirs '(L R U D)))
(loop for dir in dirs
do (let ((new_pos (move cur_pos dir)))
(if (and new_pos
(not (member new_pos *used* :test #'equalp)))
(dfs-rec new_pos (+ depth 1) cur_pos))))))
(print (dfs-rec '(0 1 2 3 4 5 6 7 8) 0 '()))
Instead of processing the four moves one by one while relying on recursion to keep track of what-to-do-next on each level, just push all the resulting positions at once to a to-do list, then pop and continue with the top one; repeating while the to-do list is not empty (i.e. there is something to do, literally):
(defun new-positions (position)
(let* ( (zero (position 0 position))
(row (floor zero 3))
(col (mod zero 3)) )
(mapcan
#'(lambda (command)
(if (cadr command)
(let ((res (copy-list position)))
(rotatef (elt res zero) (elt res (caddr command)))
(list res))))
`((L ,(> col 0) ,(- zero 1))
(R ,(< col 2) ,(+ zero 1))
(U ,(> row 0) ,(- zero 3))
(D ,(< row 2) ,(+ zero 3))) )))
; non-recursive dfs function skeleton
(defun dfs (start-pos &aux to-do curr new)
(setf to-do (list start-pos))
(loop while to-do
do (progn (setf curr (pop to-do))
(setf new (new-positions curr))
(setf to-do (nconc new to-do)))))
This way there's no more info to keep track of, with recursion -- it's all in the to-do list.
This means the generated positions will be processed in the LIFO order, i.e. the to-do list will be used as a stack, achieving the depth-first search strategy.
If you'd instead append all the new positions on each step at the end of the to-do list, it'd mean it being used as a queue, in a FIFO order, achieving the breadth-first search.
I am attempting to accept a list, have it count the positive, negative, and zeros, and return a new list.
The only thing I notice as I'm debugging is that the list is iterating through, but it is not utilizing any of the conditionals. So its successfully recursively calling itself, and then it just errors once its empty.
(define (mydisplay value)
(display value)
(newline)
#t
)
(define neg 0)
(define z 0)
(define pos 0)
(define (posneg lst)
(cond
((NULL? lst))
(NEGATIVE? (car lst) (+ 1 neg))
(ZERO? (car (lst)) (+ 1 z))
(else (+ 1 pos))
)
(posneg (cdr lst))
)
(mydisplay (posneg '(1 2 3 4 2 0 -2 3 23 -3)))
(mydisplay (posneg '(-1 2 -3 4 2 0 -2 3 -23 -3 0 0)))
(mydisplay (posneg '()))
OK, my favorite technique to apply here is wishful thinking as I learned it from Gerald Jay Sussman and Hal Abelson in the Structure and Interpretation of Computer Programs (SICP) course. Particularly, video lecture 2B. Compound Data will be of interest to you.
Let's start by just pretending (wishing) with have this data container available to us that holds 3 values: one for positives, one for negatives, and one for zeros. We'll call it pnz.
The way to create one of these is simple
; construct a pnz that has 1 positive, 4 negatives, and 2 zeros
(define x (make-pnz 1 4 2))
To select the positives value
(positives x) ;=> 1
To select a negatives value
(negatives x) ;=> 4
To select the zeros value
(zeros x) ;=> 2
Forget for a moment that these procedures don't exist (yet). Instead, just wish that they did and begin writing the procedure to solve your problem.
We'll start with some pseudocode
; pseudocode
define count-pnz xs
if xs is null? return (make-pnz p n z)
if (car xs) is positive, update the positive count by one
if (car xs) is negative, update the negative count by one
if (car xs) is zero, update the zero count by one
return count-pnz (cdr xs)
OK, that's pretty straight forward actually. Well, with one little gotcha. Notice I'm saying "update the count by one"? Well we need somewhere to store that count as the procedure is iterating. Let's make a slight adjustment to the pseudocode, this time including a pnz parameter to keep track of our current count
; pseudocode v2
define count-pnz xs pnz=(0 0 0)
if xs is null? return (make-pnz p n z)
if (car xs) is positive, nextpnz = (make-pnz p+1 n z)
if (car xs) is negative, nextpnz = (make-pnz p n+1 z)
if (car xs) is zero, nextpnz = (make-pnz p n z+1)
return count-pnz (cdr xs) nextpnz
Now this procedure makes sense to me. In the simplest case where xs is an empty list, it will simply return a pnz of (0 0 0). If xs has any number of values, it will iterate through the list, one-by-one, and increment the corresponding value in the pnz container.
Translating this into scheme is a breeze
; wishful thinking
; we will define make-pnz, positives, negatives, and zeros later
(define (count-pnz xs (pnz (make-pnz 0 0 0)))
(let [(p (positives pnz))
(n (negatives pnz))
(z (zeros pnz))]
(cond [(null? xs) pnz]
[(> (car xs) 0) (count-pnz (cdr xs) (make-pnz (+ 1 p) n z))]
[(< (car xs) 0) (count-pnz (cdr xs) (make-pnz p (+ 1 n) z))]
[(= (car xs) 0) (count-pnz (cdr xs) (make-pnz p n (+ 1 z)))])))
You'll notice I used a let here to make it easier to reference the individual p, n, z values of the current iteration. That way, when we detect a positive, negative, or zero, we can easily increment the appropriate value by simply doing (+ 1 p), (+ 1 n), or (+ 1 z) accordingly. Values that are not meant to be incremented can simply be passed on untouched.
We're getting extremely close. Our procedure make logical sense but we need to define make-pnz, positives, negatives, and zeros before it can work. By the way, this methodology of defining data objects by creating constructors and selectors to isolate use from representation is called data abstraction. You'll learn more about that in the video I linked, if you're interested.
So here's the contract that we need to fulfill
; PNZ CONTRACT
; pnz *must* behave like this
(positives (make-pnz p n z)) ⇒ p
(negatives (make-pnz p n z)) ⇒ n
(zeros (make-pnz p n z)) ⇒ z
Let's implement it !
; constructor
(define (make-pnz p n z)
(list p n z))
; positives selector
(define (positives pnz)
(car pnz))
; negatives selector
(define (negatives pnz)
(cadr pnz))
; zeros selector
(define (zeros pnz)
(caddr pnz))
Pff, well that was easy as can be ! Using a list, car, cadr, and caddr made our job simple and it's easy to understand how pnz behaves.
Without further ado, let's see this thing in action now
(define answer (count-pnz '(-1 2 -3 4 2 0 -2 3 -23 -3 0 0)))
(positives answer) ; => 4
(negatives answer) ; => 5
(zeros answer) ; => 3
And there you have it. Wishful thinking and a dash of data abstraction to the rescue.
Data abstraction is a very powerful concept. You might be thinking, "Why didn't we just use list in the count-pnz procedure instead of all of this constructor/selector ceremony?" The answer may not be readily apparent, but it is a bit too much for me to get into with this post. Instead, I sincerely do hope you check out the learning resources I linked as I'm certain they will be of great benefit to you.
#DavinTryon says "#naomik's answer could be defined in something other than a list (even just functions)."
Yep, that's totally true. Let's see make-pnz, positives, negatives, and zero implemented in a different way. Remember, the contract still has to be fulfilled in order for this implementation to be considered valid.
; constructor
(define (make-pnz p n z)
(λ (f) (f p n z)))
; selectors
(define (positives pnz)
(pnz (λ (p n z) p)))
(define (negatives pnz)
(pnz (λ (p n z) n)))
(define (zeros pnz)
(pnz (λ (p n z) z)))
Pretty cool. This demonstrates the beauty of data abstraction. We were able to completely re-implement make-pnz, positives, negatives, and zeros in a different way, but because we still fulfilled the original contract, our count-pnz function does not need to change at all.
First, let me say that #naomik's answer is awesome. This is the way to deconstruct the problem and build it up step by step.
When I first read the question, what I ended up thinking was:
How can I reduce a list of integers to the defined list '(p n z)?
So reduce means perhaps using foldl or foldr.
Here is example with foldr (returning a list in the format '(p n z)):
(define (count-pnz xs)
(foldr (lambda (next prev)
(cond ((= next 0) (list (car prev) (cadr prev) (+ 1 (caddr prev))))
((< next 0) (list (car prev) (+ 1 (cadr prev)) (caddr prev)))
(else (list (+ 1 (car prev)) (cadr prev) (caddr prev)))))
'(0 0 0) xs))
The body of the solution is the lambda we are using to reduce. Essentially, this just adds 1 to the p, n or z position of the list (using car, cadr and caddr respectively).
*note, this solution is not optimized.
A better way to keep values while computing are by making a helper that has the data you want to keep as arguments. Updating the value is the same as recursing by providing a new value:
(define (pos-neg-zero lst)
(define (helper pos neg zero lst)
(cond ((null? lst) (pnz pos neg zero)) ; the result
((positive? (car lst)) (helper (+ 1 pos) neg zero (cdr lst)))
((negative? (car lst)) (helper pos (+ neg 1) zero (cdr lst)))
(else (helper pos neg (+ zero 1) (cdr lst)))))
(helper 0 0 0 lst))
I like #naomik's abstraction but unboxing/boxing for each iteration within the helper is perhaps overkill. Though having a contract is nice and both Racket and R6RS supports making your own types:
;; scheme version (r6rs)
(define-record-type (pnz-type pnz pnz?)
(fields
(immutable p pnz-p)
(immutable n pnz-n)
(immutable z pnz-z)))
;; racket
(struct pnz (p n z) #:transparent)
An alternative would be it returned the values as separate results with values or it could take a continuation:
(define (pos-neg-zero lst . lcont)
(define cont (if (null? lcont) values (car lcont)))
(define (helper pos neg zero lst)
(cond ((null? lst) (cont pos neg zero)) ; the result
((positive? (car lst)) (helper (+ 1 pos) neg zero (cdr lst)))
((negative? (car lst)) (helper pos (+ neg 1) zero (cdr lst)))
(else (helper pos neg (+ zero 1) (cdr lst)))))
(helper 0 0 0 lst))
(pos-neg-zero '(1 -3 0)) ; returns 1, 1, and 1
(pos-neg-zero '(1 -3 0) pnz) ; returns result as an object
(pos-neg-zero '(1 -3 0) list) ; returns result as a list
(pos-neg-zero '(1 -3 0) (lambda (p n z) (+ p n z))) ; does something with the results
#!racket has optional arguments so the prototype could be just without having to have the first expression to check if there was passed an extra argument or not:
(define (pos-neg-zero lst (cont values))
...)
Here is what I have done so far:
(define sumOdd
(lambda(n)
(cond((> n 0)1)
((odd? n) (* (sumOdd n (-(* 2 n) 1)
output would look something like this:
(sumOdd 1) ==> 1
(sumOdd 4) ==> 1 + 3 + 5 + 7 ==> 16
(sumOdd 5) ==> 1 + 3 + 5 + 7 + 9 ==> 25
This is what I am trying to get it to do: find the sum of the first N odd positive integers
I can not think of a way to only add the odd numbers.
To elaborate further on the sum-odds problem, you might solve it in terms of more abstract procedures that in combination accumulates the desired answer. This isn't necessarily the easiest solution, but it is interesting and captures some more general patterns that are common when processing list structures:
; the list of integers from n to m
(define (make-numbers n m)
(if (= n m) (list n) ; the sequence m..m is (m)
(cons n ; accumulate n to
(make-numbers (+ n 1) m)))) ; the sequence n+1..m
; the list of items satisfying predicate
(define (filter pred lst)
(if (null? lst) '() ; nothing filtered is nothing
(if (pred (car lst)) ; (car lst) is satisfactory
(cons (car lst) ; accumulate item (car lst)
(filter pred (cdr lst))) ; to the filtering of rest
(filter pred (cdr lst))))) ; skip item (car lst)
; the result of combining list items with procedure
(define (build-value proc base lst)
(if (null? lst) base ; building nothing is the base
(proc (car lst) ; apply procedure to (car lst)
(build-value proc base (cdr lst))))) ; and to the building of rest
; the sum of n first odds
(define (sum-odds n)
(if (negative? n) #f ; negatives aren't defined
(build-value + ; build values with +
0 ; build with 0 in base case
(filter odd? ; filter out even numbers
(make-numbers 1 n))))) ; make numbers 1..n
Hope this answer was interesting and not too confusing.
Let's think about a couple of cases:
1) What should (sumOdd 5) return? Well, it should return 5 + 3 + 1 = 9.
2) What should (sumOdd 6) return? Well, that also returns 5 + 3 + 1 = 9.
Now, we can write this algorithm a lot of ways, but here's one way I've decided to think about it:
We're going to write a recursive function, starting at n, and counting down. If n is odd, we want to add n to our running total, and then count down by 2. Why am I counting down by 2? Because if n is odd, n - 2 is also odd. Otherwise, if n is even, I do not want to add anything. I want to make sure that I keep recursing, however, so that I get to an odd number. How do I get to the next odd number, counting down from an even number? I subtract 1. And I do this, counting down until n is <= 0. I do not want to add anything to my running total then, so I return 0. Here is what that algorithm looks like:
(define sumOdd
(lambda (n)
(cond ((<= n 0) 0)
((odd? n) (+ n (sumOdd (- n 2))))
(else (sumOdd (- n 1))))))
If it helps you, here is a more explicit example of a slightly different algorithm:
(define sumOdd
(lambda (n)
(cond ((<= n 0) 0)
((odd? n) (+ n (sumOdd (- n 1))))
((even? n) (+ 0 (sumOdd (- n 1))))))) ; note that (even? n) can be replaced by `else' (if its not odd, it is even), and that (+ 0 ..) can also be left out
EDIT:
I see that the problem has changed just a bit. To sum the first N positive odd integers, there are a couple of options.
First option: Math!
(define sumOdd (lambda (n) (* n n)))
Second option: Recursion. There are lots of ways to accomplish this. You could generate a list of 2*n and use the procedures above, for example.
You need to have 2 variables, one which keep counter of how many odd numbers are still to be added and another to hold the current odd number which gets increment by 2 after being used in addition:
(define (sum-odd n)
(define (proc current start)
(if (= current 0)
0
(+ start (proc (- current 1) (+ start 2)) )))
(proc n 1))
Here is a nice tail recursive implementation:
(define (sumOdd n)
(let summing ((total 0) (count 0) (next 1))
(cond ((= count n) total)
((odd? next) (summing (+ total next)
(+ count 1)
(+ next 1)))
(else (summing total count (+ next 1))))))
Even shorter tail-recursive version:
(define (sumOdd n)
(let loop ((sum 0) (n n) (val 1))
(if (= n 0)
sum
(loop (+ sum val) (- n 1) (+ val 2)))))