QuickSort 的迭代实现中的无限循环?
Infinite loop in iterative implementation of QuickSort?
我正在尝试使用 Lomuto's Partitiong 方法实现迭代 QuickSort,因此我正在尝试实现 stack 包含一对定义要分区的子数组的索引,使用 struct
数组和两个 fields
:iBeg
、iEnd
和 storing/accessing 只有 end
元素。
代码如下:
function [sorted] = iterativeQuickSort(A)
% Accepts 1xN unsorted integer array.
% Returns a sorted copy.
% See also Partition.
% Starting and ending indexes of unsorted array.
iBeg = 1;
iEnd = numel(A);
% Struct holding A's subarrays star/end indexes resulting from partitioning.
stack_elem = struct('iBeg', iBeg, 'iEnd', iEnd);
stack(end + 1) = stack_elem; % push on stack
while numel(stack) != 0
% Extract last pair of indexes.
iBeg = stack(end).iBeg;
iEnd = stack(end).iEnd;
stack(end) = []; % pop from stack
% Get pivot index and array after rearranging elements around the pivot.
[B, pivotIndex] = Partition(A, iBeg, iEnd);
A = B;
% Store indexes of the next two subarrays defined by the pivot index,
% if their sizes are > 0.
if pivotIndex - 1 > iBeg
stack_elem = struct('iBeg', iBeg, 'iEnd', pivotIndex - 1);
stack(end + 1) = stack_elem;
end
if pivotIndex + 1 < iEnd
stack_elem = struct('iBeg', pivotIndex + 1, 'iEnd', iEnd);
stack(end + 1) = stack_elem;
end
end
sorted = A;
end
function [A, pivotIndex] = Partition (A, iBeg, iEnd)
% Accepts 1xN integer array.
% Two integers - start and end indexes current subarray of A.
% Returns index of pivot element of current subarray partition
% and A after swaps.
pivotValue = A(iEnd); % Choose last element to be pivot.
pivotIndex = iBeg; % Initialize pivot index to start of subarray.
for i = iBeg : iEnd % Iterate over current subarray
if A(i) <= pivotValue % Push elements <= pivot in front of pivot index.
% Place element at i-th position before element with pivot index.
[A(i), A(pivotIndex)] = swapElements(A(pivotIndex), A(i));
% Account for the swap, go to next element.
pivotIndex = pivotIndex + 1;
end
end
% Bring the element used as pivot to its place
[A(iEnd), A(pivotIndex)] = swapElements(A(pivotIndex), A(iEnd));
end
function [elem2, elem1] = swapElements(elem1, elem2)
[elem2, elem1] = deal(elem1, elem2);
end
明显愚蠢的数组赋值 A = B
是为了指示由于 swaps
引起的元素更改在函数 Partition(A, iBeg, iEnd)
.
执行后保留
目前的状态似乎是一个无限循环,其原因我无法确定,因此将不胜感激任何建议和建议!
输入:
A = [5, 4, 6, 2, 9, 1, 7, 3];
S = iterativeQuickSort(A)
预期输出:
[1, 2, 3, 4, 5, 6, 7, 9]
当前输出:从不 returns 函数,仅通过强制制动停止:ctrl + c。
注意:分区函数的实现和应用与指出可能重复的不同。
您的 Partition
函数有错误。如果您查看维基百科页面上的伪代码,您会看到循环条件是:
for j := lo to hi - 1 do
你的Partition
函数中对应的行是:
for i = iBeg : iEnd % Iterate over current subarray
通过转到 iEnd
而不是 iEnd - 1
,您将枢轴值与其自身进行比较,并以差一的枢轴索引结束。只需将此行更改为:
for i = iBeg : iEnd-1 % Iterate over current subarray
我正在尝试使用 Lomuto's Partitiong 方法实现迭代 QuickSort,因此我正在尝试实现 stack 包含一对定义要分区的子数组的索引,使用 struct
数组和两个 fields
:iBeg
、iEnd
和 storing/accessing 只有 end
元素。
代码如下:
function [sorted] = iterativeQuickSort(A)
% Accepts 1xN unsorted integer array.
% Returns a sorted copy.
% See also Partition.
% Starting and ending indexes of unsorted array.
iBeg = 1;
iEnd = numel(A);
% Struct holding A's subarrays star/end indexes resulting from partitioning.
stack_elem = struct('iBeg', iBeg, 'iEnd', iEnd);
stack(end + 1) = stack_elem; % push on stack
while numel(stack) != 0
% Extract last pair of indexes.
iBeg = stack(end).iBeg;
iEnd = stack(end).iEnd;
stack(end) = []; % pop from stack
% Get pivot index and array after rearranging elements around the pivot.
[B, pivotIndex] = Partition(A, iBeg, iEnd);
A = B;
% Store indexes of the next two subarrays defined by the pivot index,
% if their sizes are > 0.
if pivotIndex - 1 > iBeg
stack_elem = struct('iBeg', iBeg, 'iEnd', pivotIndex - 1);
stack(end + 1) = stack_elem;
end
if pivotIndex + 1 < iEnd
stack_elem = struct('iBeg', pivotIndex + 1, 'iEnd', iEnd);
stack(end + 1) = stack_elem;
end
end
sorted = A;
end
function [A, pivotIndex] = Partition (A, iBeg, iEnd)
% Accepts 1xN integer array.
% Two integers - start and end indexes current subarray of A.
% Returns index of pivot element of current subarray partition
% and A after swaps.
pivotValue = A(iEnd); % Choose last element to be pivot.
pivotIndex = iBeg; % Initialize pivot index to start of subarray.
for i = iBeg : iEnd % Iterate over current subarray
if A(i) <= pivotValue % Push elements <= pivot in front of pivot index.
% Place element at i-th position before element with pivot index.
[A(i), A(pivotIndex)] = swapElements(A(pivotIndex), A(i));
% Account for the swap, go to next element.
pivotIndex = pivotIndex + 1;
end
end
% Bring the element used as pivot to its place
[A(iEnd), A(pivotIndex)] = swapElements(A(pivotIndex), A(iEnd));
end
function [elem2, elem1] = swapElements(elem1, elem2)
[elem2, elem1] = deal(elem1, elem2);
end
明显愚蠢的数组赋值 A = B
是为了指示由于 swaps
引起的元素更改在函数 Partition(A, iBeg, iEnd)
.
目前的状态似乎是一个无限循环,其原因我无法确定,因此将不胜感激任何建议和建议!
输入:
A = [5, 4, 6, 2, 9, 1, 7, 3];
S = iterativeQuickSort(A)
预期输出:
[1, 2, 3, 4, 5, 6, 7, 9]
当前输出:从不 returns 函数,仅通过强制制动停止:ctrl + c。
注意:分区函数的实现和应用与指出可能重复的不同。
您的 Partition
函数有错误。如果您查看维基百科页面上的伪代码,您会看到循环条件是:
for j := lo to hi - 1 do
你的Partition
函数中对应的行是:
for i = iBeg : iEnd % Iterate over current subarray
通过转到 iEnd
而不是 iEnd - 1
,您将枢轴值与其自身进行比较,并以差一的枢轴索引结束。只需将此行更改为:
for i = iBeg : iEnd-1 % Iterate over current subarray