Description
Problem 1. The nth Tetranacci number Tn is mathematically defined as follows: T0 = 0 T1 = 1 T2 = 1 T3 = 2 Tn = Tn−1 + Tn−2 + Tn−3 + Tn−4. Note carefully that the inductive (or recursive) part in this definition returns the value of Tn based on the four preceding values {Tn−1, Tn−2, Tn−3, Tn−4}. Your task is to translate the preceding definition into a recursive OCaml function tetra1 : int -> int which accepts an integer n (you may assume that n ≥ 0), and computes the n-th Tetranacci number (don’t worry about efficiency, only about making your OCaml code as close as possible to the mathematical definition of Tn). Ungraded Bonus: 1 What is the approximate number of times this function will be called on an input of size n? Express this approximate number as a function of n. Problem 2. You will now implement a different and more efficient version of the function tetra1, called tetra2 : int -> int, for computing Tetranacci numbers. (a) Implement an OCaml function sum top 4 : int list -> int list, which takes a list of integers and adds the sum of the top four elements to the head of the list (e.g. the input list [1;1;1;1;3;2] should become the output list [4;1;1;1;1;3;2]). (b) Implement a recursive OCaml function ascending : int -> int list, which accepts an integer n as input (again, assume n ≥ 0), and returns a list of integers from 0 to n in ascending order. 1You should be able to answer this question if you took, or are now taking, the course CS 330 “Introduction to Analysis of Algorithms” (or an equivalent course at another university). 1 (c) Implement a recursive OCaml function tetra2 : int -> int, which computes the nth Tetranacci number in “time linear in n”, i.e. it computes Tn by making one recursive call (instead of 4 recursive calls as it was for function tetra1 in Problem 1). Put differently, the number of recursive calls that tetra2 makes is linear in n. Problem 3. The Fibonacci k-step numbers are a generalization of the Fibonacci and Tetranacci sequences, where Fi = 0 for i ≤ 0, F1 = 1, F2 = 1, and for every j ≥ 3: Fj = Fj−1 + Fj−2 + · · · + Fj−k, i.e. Fj is the sum of the k previous numbers in the sequence. You will implement an OCaml function fib k step : int -> int -> int for computing these numbers efficiently. (a) Implement an OCaml function sum top k : int -> int list -> int, which accepts an integer k along with a list of integers, and sums up the first k elements of the list. If the list has length less than k, simply sum all the elements in the list (assume that an empty list evaluates to a sum of 0). (b) Implement an OCaml function fib k step : int -> int -> int, which accepts two integers n and k and, for a fixed k, computes in “time linear in n” the nth Fibonacci k-step number.2 For the three next problems you will deal with the data type of polymorphic binary trees. The solution to each part is meant to be very short (one to four lines). You may (and should) use functions from earlier parts to solve later parts. Unless otherwise specified, your solutions may utilize functions from the standard OCaml library as well as material from lecture notes. Problem 4. Define a polymorphic data type called ’a binTree as follows: type ’a binTree = | Leaf | Node of ’a * (’a binTree) * (’a binTree) Trees of this type will contain a single piece of data of type ’a at each node, and no data at their leaves. (a) Define an OCaml function mapT : (’a -> ’b) -> ’a binTree -> ’b binTree which operates on trees just as map operates on lists. In other words, mapT takes a function and applies it to every data item of type ’a in a tree of type ’a binTree. 2Hint: Do part (c) of Problem 2 before attempting part (b) of Problem 3; the latter generalizes the former. (b) Define an OCaml function foldT : (’a -> ’b -> ’b -> ’b) -> ’a binTree -> ’b -> ’b which operates on trees the same way that foldr operates on lists. In other words, the base item of type ’b should replace the leaf constructor in the tree, and the function of type ’a -> ’b -> ’b -> ’b should replace the node constructor in the tree.3 Problem 5. In each of the three parts in this problem, you will get full credit if you use foldT and define at most one helper function. (a) Define an OCaml function leafCount : ’a binTree -> int which returns an integer representing the total number of leaves in a tree. (b) Define an OCaml function nodeCount : ’a binTree -> int which returns an integer representing the total number of non-leaf nodes in a tree. (c) Define a function height : ’a binTree -> int which returns an integer representing the height of the tree. Trees consisting of only a leaf have height 0. Problem 6. (a) A tree is perfect if all the leaves of the tree are at the same depth. Define an OCaml function perfect : ’a binTree -> bool which returns true if the tree supplied is perfect and false otherwise. You may use any approach to implement this function. (b) A tree is a degenerate if all the nodes are arranged in a single path. Equivalently, a tree is degenerate if all nodes have at least one leaf child. Define an OCaml function degenerate : ’a binTree -> bool which returns true if and only if the tree supplied is degenerate.4 (c) Define a function treeToList (t : ’a binTree) : (’a list) option. If the supplied tree is degenerate, the function should return Some l, where l corresponds to a list constructed by replacing the Node constructors with (::) constructors and replacing the Leaf constructors with the [] constructor where appropriate. If the supplied tree is not degenerate, the function should return None. You are encouraged to use degenerate and foldT to implement this function.5 3Hint: It is useful to recall how the function foldr right operates on lists. If you ask for the type of foldr right at the top level, you get the following: # List.fold right ;; – : (’a -> ’b -> ’b) -> ’a list -> ’b -> ’b = Try to understand how foldr right works before attempting to program foldT. AAA 4Hint: You do not need to make your function recursive if you use functions you have already defined. How many leaves does a degenerate tree have? How many nodes? 5Hint: Do not make your use of foldT more complicated than necessary. Do you need to check that the tree is degenerate more than once? Use your nodeCount implementation as a guide.
