Strand sort is a recursive sorting algorithm that sorts items of a list into increasing order. It has O(n2) worst time complexity which occurs when the input list is reverse sorted.[1] It has a best case time complexity of O(n) which occurs when the input is a list that is already sorted.
The algorithm first moves the first element of a list into a sub-list. It then compares the last element in the sub-list to each subsequent element in the original list. Once there is an element in the original list that is greater than the last element in the sub-list, the element is removed from the original list and added to the sub-list. This process continues until the last element in the sub-list is compared to the remaining elements in the original list. The sub-list is then merged into a new list. Repeat this process and merge all sub-lists until all elements are sorted. This algorithm is called strand sort because there are strands of sorted elements within the unsorted elements that are removed one at a time. This algorithm is also used in J Sort for fewer than 40 elements.[2]
This example is based on the description of the algorithm provided in the book, IT Enabled Practices and Emerging Management Paradigms.
Step 1: Start with a list of numbers:
Step 2: Next move the first element of the list into a new sub-list: sub-list contains
Step 3: Then iterate through the original list and compare each number to 5 until there is a number greater than 5.
Step 4: Now compare 9 with the remaining elements in the original list until there is a number greater than 9.
Step 5: Now there are no more elements to compare 9 to so merge the sub-list into a new list, called solution-list.
After step 5, the original list contains
The sub-list is empty, and the solution list contains
Step 6: Move the first element of the original list into sub-list: sub-list contains
Step 7: Iterate through the original list and compare each number to 1 until there is a number greater than 1.
Step 8: Now compare 4 with the remaining elements in the original list until there is a number greater than 4.
Step 9: Now compare 6 with the remaining elements in the original list until there is a number greater than 6.
Step 10: Now compare 8 with the remaining elements in the original list until there is a number greater than 8.
Step 11: Since there are no more elements in the original list to compare to, the sub-list is merged with the solution list. Now the original list contains, the sub-list is empty and the solution-list contains: .
Step 12: Move the first element of the original list into sub-list. Sub-list contains
Step 13: Iterate through the original list and compare each number to 2 until there is a number greater than 2.
Step 14: Now compare 3 with the remaining elements in the original list until there is a number greater than 3.
Step 15: Since there are no more elements in the original list to compare to, the sub-list is merged with the solution list. The original list now contains, the sub-list is empty, and solution list contains: .
Step 16: Move the first element of the original list into sub-list. Sub-list contains .
Step 17: Since the original list is now empty, the sub-list is merged with the solution list. The solution list now contains: . There are now no more elements in the original list, and all of the elements in the solution list have successfully been sorted into increasing numerical order.
Since Strand Sort requires many insertions and deletions, it is best to use a linked list when implementing the algorithm.[3] Linked lists require constant time for both insertions and removals of elements using iterators. The time to traverse through the linked list is directly related to the input size of the list.[4] The following implementation is done in Java 8 and is based on the description of the algorithm from the book, IT Enabled Practices and Emerging Management Paradigms.
import java.util.*;
public class strandSort