Measuring algorithm performance

Measure Algorithm Performance
By: Asfaw Alene &Habitamu Asimare
Bahir Dar
Institute of Technology, BiT
Algorithm Analysis
Submitted to: Dr Million M.
06/20/2018

Outline
1. Overview
2. Asymptotic Notations
3. Analysis
 Average
 Best
Worst-Case
4. Concluding
5. References

1. Overview of Algorithm
An algorithm is a step by step procedure for solving a problem, based
on conducting a sequence of specified actions.
A computer program can be viewed as an elaborate algorithm. In
mathematics and computer science, an algorithm usually means a finite
procedure that solves a real world problem.
An algorithm can be used in
For search engines
Encryption technique
Memory management
Resource allocation (Operating Systems implementations)

Continued…
Basically Algorithm plays a great role to make computers system efficiently. The
following is the points to be remembered in Algorithm
 An algorithm is a sequence of unambiguous instructions.
 An algorithm is a well-defined procedure that allows a computer to solve a
problem
 The algorithm is described as a series of logical steps in a language that is
easily understood
 Algorithms is computer understandable actions that can be implemented in
Programming languages
 In fact, it is difficult to think of a task performed by your computer that does
not use algorithms.

2. Asymptotic Notations
Three asymptotic notations are mostly used to represent time
complexity of algorithms.
1. The theta(Θ) Notation
 bounds a functions from above and below, so it defines exact asymptotic behavior.
 A simple way to get Theta notation of an expression is to drop low order terms and ignore leading
constants.
 For example, consider the following expression.
3n3 + 6n2 + 6000 = Θ(n3)
2. Big O Notation
 defines an upper bound of an algorithm, it bounds a function only from above.
 have to use two statements for best and worst cases in theta notations:
1. The worst case time complexity of Insertion Sort is Θ(n^2).
2. The best case time complexity of Insertion Sort is Θ(n).
 In case of Big O notations we select worst case which is O(n^2)
3. Omega(Ω) Notation
 Just as Big O notation provides an asymptotic upper bound on a function, Ω notation provides an
asymptotic lower bound
 the Omega notation is the least used notation among all three.

3. ANALYSIS
We can have three cases to analyze an algorithm:
 Worst Case
In the worst case analysis, we calculate upper bound on running time of an algorithm.
We must know the case that causes maximum number of operations to be executed.
 Average Case
n average case analysis, we take all possible inputs and calculate computing time for all of the
inputs.
Sum all the calculated values and divide the sum by total number of inputs. We must know (or
predict) distribution of cases.
 Best Case
 we calculate lower bound on running time of an algorithm. We must know the case that causes minimum
number of operations to be executed.

Asymptotic notations with examples
1 < logn < 𝑛 < n < nlogn< n < n2<n3 < ……. 2n < 3n < .. nn
Lower bound Upper bound
Asymptotic notations are mathematical tools to represent time complexity of
algorithms for asymptotic analysis.
The mostly used asymptotic notations
1. Big O Notation:
2. Θ Notation
3. Ω Notation

Big O-Notation (Upper limit )
For the function F(n) = O(n) , iff there exists positive constants c and n0 ,
such that f(n) ≤ c* g(n) for all n ≥ n0.
Example f(n)= 2n+3; what is the time complexity of F(n) ?
Solution chose the c value and g(n) value on the since O
notation is the upper bound so it is only possible to chose the
upper value. so we can chose c=10, and g(n)=n , for n>=1
• 2n+3 <10n is true
the time complexity of F(n) will be = O(n).

Omega Ω Notation (lower bound )
For the function F(n) = Ω( n) , iff there exists positive constants
C and n0 , such that f(n) ≥ c* g(n) for all n ≥ n0.
Solution chose the c value and g(n) value on the since Ω
upper value. so we can chose c=1, and g(n)=n , for n>=1
2n+3 > n is true
the time complexity of F(n) will be = Ω (n).

Omega Ω Notation (lower bound )
For the function F(n) = Ω( n) , iff there exists positive constants
C and n0 , such that f(n) ≥ c* g(n) for all n ≥ n0.
Solution chose the c value and g(n) value on the since Ω
upper value. so we can chose c=1, and g(n)=logn , for n>=1
2n+3 > logn is true
the time complexity of F(n) will be = Ω (logn).

Theta Θ-Notation (contd)
For the function F(n) = Θ(g(n)) , iff there exists positive constants c1, c2
and n0 , such that c1*gn <= f(n) ≤ c2* g(n) for all n ≥ n0. }
Solution chose the c1,c2 value and g(n) value on the since Θ
notation is the average bound so it is only possible to chose the
value which is the multiple of n. so we can chose c1=1, c2 = 5,
and g(n)=n , for n>=1
• 1n< 2n+1<5n is true, so
the time complexity of F(n) will be = Θ(n).

BEST CASE ANALYSIS (CONTD …)
Consider for the following example
A linear searching
From this example if we are searching a key element that are
present at first index is called best case
Best case time will be = 1, i.e B(n) or we can use notations
Big-O Notation and will be O(n)=1;
2
8 6 12 5 7 9 4 3 16

WORST CASE ANALYSIS
In the worst case analysis, we calculate upper bound on
running time of an algorithm. We must know the case that
causes maximum number of operations to be executed.
A linear searching worst case analysis
From this example if we are searching a key element that are
present at last of the index elements are called best case
2
8 6 12 5 7 9 4 3 16

AVERAGE CASE ANALYSIS
A linear searching average case analysis
 From this example if we are searching a key element that are
present at middle of the index elements are called best case.
If we are looking from n elements will be
A(n)= (n+1)/2
What if you are given to analyze the case’s in Binary Search Tree ?
8 6 12 5 7 9 4 3 16

ADVANTAGE AND DISADVANTAGES
Advantages Disadvantages
1. To understand the complexity
that the program require
1. does not account for memory access times
1. to minimize unnecessary
operations
1. data size can determines the algorithm
behavior
1. solve real-world problem 1. sometimes the best and worst cases boundary
can’t algorithm performance

4. Conclusion
Algorithms are the backbone of the working structure of computer
system.
Effectiveness and Correctness allows us to analyze the performance of
algorithm.
Beyond working on algorithm we have two think of two major aspects
of algorithm
Time :
 Instructions take time.
 How fast does the algorithm perform?
 What affects its runtime?
Space :
 Data structures take space
 What kind of data structures can be used?
 How does choice of data structure affect the runtime?
 So space and time is challenges of algorithm performance

Continued…
Strength of measuring algorithm performance analysis
 It expresses the amount of work for a given algorithm as being
proportional to a bounding function.
Weakness of measuring algorithm performance analysis
independent of implementation details such as the choice of computer
hardware
 The asymptotic notation can be equal in some cases, like worst and best
cases.
SO that when the performance of algorithm is done, the analysis
should include what type of the resource should support while
implementing/coding the algorithm

5. Reference
• Weiss, M.A., Data Structures and Problem Solving Using JAVA, Third Edition,
AddisonWesley Publishing Company, Inc., Reading, MA, 2006, pp. 313-331.
• Cormen, T. H., Leiserson, C. E., Rivest, R. L., & Stein, C. (2009). Introduction to
algorithms (3rd ed). MIT Press
• Musselman, R. (2007). Robustness: A better measure of algorithm performance.
Thesis, Naval Postgraduate School, Monterey, CA.
• McMaster, K., Sambasivam, S., Rague, B., & Wolthuis, S. (2015). Distribution of
Execution Times for Sorting Algorithms Implemented in Java. Proceedings of
Informing Science & IT Education Conference (InSITE) 2015, 269- 283
• Mr. Rajinikanth B., What is Performance Analysis of an algorithm? , 2009,
http://btechsmartclass.com/DS/U1_T2.html last accessed on 17 June 2018.

Measuring algorithm performance

Measuring algorithm performance

More Related Content

What's hot

Similar to Measuring algorithm performance

Recently uploaded

Measuring algorithm performance