Automatic Test Data Generation for Big Data Input Values Using a Genetic Algorithm and Binary Search

Abstract: Software program analysis is the process of analyzing the component of software with the several type of test situation as a way to develop quality of assess and defects. This documentation concluded that in a program coverage of path technique of GA (Genetic algorithm) hybrid with BS (Binary Search) algorithm in the values of Big Data Input terms for achieving testing of software. Automation of software analysis increases the efficiency and effectiveness of analysis. Technique of BSGA (Binary Search and Genetic Algorithm) are implement: in this procedure reeiingig the test of software efficiency.by removing the bugs prone path   from a program. BSGA provides a robust non  linear techniques of search and solution of better quality.so reduction in software testing industry by select  optimal sub domain  of big dats input values and

The experiments perform results are hold out to calculate the effects of the proposed BSGA compared to the techniques of GA .in step to find the optimal test cases and test data to input Big Data domain amounts. Whereas  ,this output minimize the test cost.  

 Keywords; Genetic Algorithm, Binary Search, Big Data ,Path coverage, Test data ,Software Testing

1.      introduction;

Software engineering(SE)is a systematic approach that maintain  the engineering methods to discipline  the process of  software’s development and thus providing authentic work of software on real atmosphere .SE encompasses activities set, namely engineering requirements, designs ,implementation ,testing and management  .

[1] while one of the most important activities is testing that effects the software development cost .huge   intensive software system of a computer has fastly developed and in the next decades tendency will be continued .These Software System handling the huge amount of data or data complexity lending to term of Big Data (BD).further in the processing of data traditional approach is to be deal insufficiently.

Holding data, data investigates   , data transfer, visualization, search, developing privacy of information and storage of data are main hurdles in BD.

[2]The BD’s significant is   that it is huge to the point it can’t be practically devoured by any framework in a span genuinely sufficient short to prepare a significant expectations.

One of the BD characteristics is volume. Which deal’s about scale of data. The BD’s primary attribute can also be quantified by counting records, files, tables or transactions.

If you focus on the world virtually all the industries including  banking automobile, avionics, telecommunications, hospitals, pharmacy oil and many other industries based to produce BD and the software implemented in these estate dealing a huge amount of input data.

[2] the computers are working by the compound of hardware and software. The software is clarifying as to alter a separate set of inputs to a obscure set of outputs in a limited period of time.

software system testing  is a procedure to usual to facilitate and developed the correction, completion, reliability, quality and security of a developed software system .the possible cost  saving from control software bugs  at interim of a software development cycle, rather than the other cycles that has been predictable  at exactly forty billions bucks by the NIST(National Institute of Standards and Technology).this amount  highlight that the present steps of testing is equivalent hence, depletion of software errors and bugs may be an essential study with a considerable income.

The large empire range (egad- {ai…an}) in BD creates more powerful relationship between the number of input data for every variable.

This paper launch a BSGA (Binary Search Genetic Algorithm) that is referred by the dependencies of data flow to look or test analyze the data satisfy a standout between the most requesting in the  classification  of data  brook way on    resolve basis.

The search of GA leads by creating a new data test from recently evaluated test data successfully.

In characterization of every chromosome(variable), the proposed skills reduce  the size of chromosome by decreasing the amount of input dates using Binary Search.

This paper is organized as;

SEC1   establish the need of GA, B Coverage of testing and to be covered the research area.

SEC 2; manifest the literature analysis.

SEC 3 present the flow of data review. for example, program

SEC4 shows the proposed approach.

SEC 5 presents the implementation of the proposed algorithm with the help of examples of precisions of small as well as large.

SEC 5also analysis the results and evaluating the efficiency by finding the comparison.

At last,

SEC 6 presents the conclusion and future prospects of this algorithm.

2. Literature   Review;

Conditionally, software testing design phase explains the amount of test cases; moreover, and based on the case of test generation strategy .however it’s a simple view  to estimate the quantity of test cases instead o calculating the  test cases number

{8}{9}

This status achieved the attention as old as the work of {10} the demanded   test cases of a system which estimate the quantity by the developed tools.in which this amount represents the quantity of necessary path of the whole program to satisfy the criterion .somehow ,this work used the counting approach of the way that are possible and are generated by  using the property of  backtracking of Prolog programming language.

An aimed program which is to be tested is to be divided into a numbers and blocks and are to be run through while processing the program. Commonly, to maintain the path of test and estimate the quantity of test cases .this approach can be followed;{11}{8}{9};

1.      Define function points(FPs).

2.      Divide the type of test (Functional, Performance/Stress, Interoperability, security etc ….)is to be used.

3.      Map the function points to the testing types(FPs)

4.      For every FP, test techniques are specified(BVAEP), Tables Decision, State Transition etc

5.      Now a maintained table, which contains a rough estimation of the quantity of test cases for every FP having it’s testing techniques (FP vs Testing Techniques vs Estimated No. Of Test Cases).

6.      Lastly, make a difference between the data of this table with Historical Data.

Figure 1: Genetic Algorithm (GA) based method to select a set of test cases proposed by (Whitten 1998)

Where every FP represents a node and the approach in sequence at the edge between these pair of nodes (FPs).

This approach creates many ways of test especially the reasonable cost of this software.

Such situations required time to increase the performance of testing.it has been confirmed that the selecting the optimal path of test could be a wise solution for decreasing the time of testing. While accepted the coverage of percentage, selecting best way to classified a optimization problem.

An example of the efforts for calculating test is the method of synthesis which utilizes the calculus refinement to define abstraction rules.which are used to calculate later  test case scenarios from project’s requirements.{12}

Discuss on GA(Genetic Algorithm),a procedure that uses GA to select a set test cases  that are to be used to test a software program which is illustrated  byFig1,is defined  by {13}.well, this procedure has implicit challenges  that were revealed later as the work by the other researchers which presents the GC(Generalized Crossing)procedure as a solution for optimizing the block sequences  and best way to  be discovered .this GC procedure, which presented by {14},iteratively  procedure of the backtracking, the author of this task claimed that this procedure performed better then GA based one.

The proposed approach in {15}represents the automatic test data  generation technique which uses a GA ,which is guided  by the data flow dependencies in the program to cover the Def User  for the search data .New test data from previously generated test data that are evaluated as effectiveness of test data and search by constructing to the conducts to the GA.for programs test data generations are used to approach with /without procedures and loops