N. Gupta, A. P. Mathur, and M. L. Soffa. Automated test data generation using an iterative relaxation method. In SIGSOFT '98/FSE-6: Proceedings of the 6th ACM SIGSOFT international symposium on Foundations of software engineering, pages 231-- 244. ACM Press, 1998.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....expected outputs. This makes the execution of a huge number of test cases impractical. 5 Related Work Our work is basically an extension of the work of Chays et al. 1] There are very few work whose targets are close to testing of database applications. Constraint solving techniques (e.g. 5] [6]) transform a procedure into a constraint system and then solve the system. Some other work are mainly for evaluating the performance of DBMS. For example, Gray et al. 7] considered how to generate huge synthetic databases by filling dummy information. Their work is not useful for us because we ....

N. Gupta, A. Mathur, M. L. Soffa. Automated test data generation using an iterative relaxation method. In Proc. of Foundations of Software Engineering, ACM Press, Nov 1998.

....and goal oriented test data generation. Random test data generation [BM83,DN84] consist in trying test data generated randomly until the selected statement is reached. This approach is not appropriate for programs with float variables as the search space is large. Path oriented approaches [OJP97,GMS98] attack the problem by first selecting a set of paths that covers all the statements satisfying a given criterion and then generating a test input which executes each selected path. These approaches include, among others, symbolic evaluation and program execution based test data generation. ....

....a path is traversed. The constraints are then symbolically solved. Although symbolic evaluation is promising, it still has several weaknesses including the handling of arrays, indeterminate loops, dynamic data structures and the size of symbolic expressions. In the program execution based approach [GMS98,GMS99], a first test data is initiated with a (randomly) chosen input. This input is then iteratively refined, by execution of the program, to obtain a final input executing the path. This approach exploits its dynamic nature to overcome some limitations of the approaches based on symbolic evaluation. ....

[Article contains additional citation context not shown here]

Neelam Gupta, Aditya P. Mathur, and Mary Lou Soffa. Automated test data generation using an iterative relaxation method. In ACM SIGSOFT Sixth International Symposium on Foundations of Software Engineering(FSE-6), pages 231--244, Orlando, Florida, November 1998.

....node, the flow at an earlier point may accidently change. Actual execution also su#ers from the speed of execution for the program to analyze. Besides, to monitor the program flow code is instrumented, i.e. to put probes in the program to ascertain path traversal. In an article by Gupta et al. [7] a hybrid of the two forms is presented. It combines the gains of both kinds, thus it does not requires as many executions to find an appropriate input. 3.1.2. Random Test Data Generation Random testing is the simplest method of generation techniques. It could actually be used to generate input ....

....involves solving a system of equations. If the system has no solution we can conclude that the path given is indeed infeasible. The problem is that solving an arbitrary system of equations is undecidable. If the system is linear we can by Gaussian elimination conclude whether that path is feasible [7]. For non linear systems it becomes more inconvenient. All methods studied have set a highest number of iterations to do before abandoning the path as infeasible in order not to end up in an infinite loop. 4.6. Constraint Satisfaction All encountered methods (except for random testing) have to ....

[Article contains additional citation context not shown here]

N. Gupta, A. P. Mathur, and M. L. So#a. Automated test data generation using an iterative relaxation method. In Proceedings of the ACM SIGSOFT sixth international symposium on Foundations of software engineering, pages 231--244, November 1998.

....may require many iterations before a suitable input is found. Upon changing the flow at a particular point, the flow at an earlier stage may accidently change. Additional information regarding automatic test data generation can be found in a survey by Edvardsson [3] In an article by Gupta et al. [6] a hybrid between the static and dynamic approach is presented. It combines the gains of both kinds, thus it does not require as many executions to find an appropriate input. Instead of doing complex symbolic or actual executions of the program, they propose that a linear representation of each ....

N. Gupta, A. P. Mathur, and M. L. So#a. Automated test data generation using an iterative relaxation method. In Proceedings of the ACM SIGSOFT sixth international symposium on Foundations of software engineering, pages 231--244, November 1998.

....programs. In general, their approach consists of a branch predicate collector, which derives a system of linear inequalities representing the branch predicates for a given path in the program. This system is solved using a solving technique of theirs called the Unified Numerical Approach (una) [5, 7]. In this paper we show that in contrast to traditional optimization methods the una is not bounded by the size of the solved system. Instead it depends on how input is composed. That is, even for very simple systems consisting of one variable we can easily get more than a thousand iterations. We ....

....This system is then solved using Gaussian elimination [11] Often, however, this technique su#ered from the fact that the derived equation system was over or under determined. As a result, Gupta et al. developed a new constraint solving technique known as the Unified Numerical Approach (una) [5, 7]. Unfortunately, una has the following weaknesses: a) There are circumstances where una fails to find a solution to a mixed integer problem, given that there is at least one. b) The number of iterations used by una to find a real value solution has a worst case bounded by infinity for all ....

[Article contains additional citation context not shown here]

N. Gupta. Automated Test Data Generation Using Iterative Relaxation Methods. PhD thesis, Purdue University, West Lafayette, Indiana, December 1999.

....the problem of choosing appropriate test data. In general, choosing input data to follow a path through the reachability graph in such a way that it correlates with the values of modeled variables is undecidable. In practice, it may be possible to use symbolic execution [5] or some heuristics [10, 14] or to use random test data, aborting those executions that are not exploring the part of the reachability graph of that is of interest. In this section we propose an approach for choosing values of variables that are modeled by the veri er. Many FSV approaches are capable of modeling system ....

N. Gupta, A. Mathur, and M. L. Soa. Automated test data generation using an iterative relaxation method. In Proceedings Foundations of Software Engineering. ACM Press, Nov. 1998.

....with the derived constraints by the constraint solver. The key step of the above process is the derivation of the system of linear constraints from the program. We discuss two approaches to carry out this task: one is based upon symbolic evlauation [3] and the other is an execution based [24] approach. We illustrate these approaches through an example program of Fig. 11 which reads three input values into variables a, b and c and then changes the contents of the variables such that eventually a contains the smallest number and c contains the largest number. We consider the generation ....

N. Gupta, A.P. Mathur, and M.L. Soa, \Automated Test Data Generation Using an Iterative Relaxation Method", ACM SIGSOFT Sixth International Symposium on Foundations of Software Engineering, pages 231-244, November 1998.

No context found.

Neelam Gupta, Aditya P. Mathur, and Mary Lou Soa, \Automated Test Data Generation using An Iterative Relaxation Method" ACM SIGSOFT Sixth International Symposium on Foundations of Software Engine ering(FSE-6), pages 231-244, Orlando, Florida, November 1998. 5

....There are a number of approaches that can be used to generate data that executes a branch. One approach is essentially to use the same technique as path coverage in that a particular path is chosen that goes through the selected branch and an attempt is made to generate test data for that path [11, 8, 1, 6]. If the effort is unsuccessful, another path is chosen, and the process continues until the branch is exercised or no more paths can be tried. The major drawback of this approach is the impact of infeasible paths. If an infeasible path is selected, then a significant computational effort could be ....

....paths in the program, path selection is an important problem in generating test data to exercise the test branch. Different paths can offer different resistance to a test data generation technique that attempts to force execution through them. We presented an iterative relaxation technique in [8] to generate test data for a given path. The new idea here is how we use the approach presented in [8] to guide the path selection process by dynamically switching execution to a path that offers less resistance to force execution to reach the given branch. Our technique starts with a given test ....

[Article contains additional citation context not shown here]

Neelam Gupta, Aditya P. Mathur, and Mary Lou Soffa, "Automated Test Data Generation using An Iterative Relaxation Method," ACM SIGSOFT Sixth International Symposium on Foundations of Software Engineering(FSE-6), pages 231-244, Orlando, Florida, November 1998.

....of basic type int or f loat as specified in the type definition of the argument. Therefore, for our current evaluation using scientific programs, it suffices not to consider test data generation for pointer inputs that point to data structures such as linked lists, trees etc. Please refer to [6] for a description of how our technique can be used to generate test data for functions with input parameters that are records or pointers to dynamic data strucures. 5 Evaluation We selected functions from the [9] for experimental evaluation of our technique. The functions have numeric inputs ....

Neelam Gupta, "Automated Test Data Generation Using Iterative Relaxation Methods." PhD. Thesis, Department of Computer Sciences, Purdue University, West Lafayette, Indiana, August 1999 (expected) .

....the desired outcome. These methods are inefficient in handling infeasibile paths. If the selected path is infeasible and the infeasibility is not detected, a large number of attempts can be performed before the search procedure terminates and a lot of effort can be wasted [7] In a previous paper [5], we proposed a method for automatically generating test data using an iterative relaxation technique. In this method, test data generation is initiated with an arbitrarily chosen input from a given domain. If the given path is not traversed on this input, a set of linear constraints on increments ....

....are added to the input to obtain a new input. If the path is traversed on the new input then the method terminates. Otherwise, the steps of refining the input are carried out iteratively using a relaxation technique to obtain the desired input. In this paper, we improve on our previous technique [5] and also present experimental results demonstrating that the new technique is viable, practical and effective. An important step in our previous method is solving the set of linear inequalities obtained from the branch predicates and then using the solution to refine the current input. If the ....

[Article contains additional citation context not shown here]

N. Gupta, A. P. Mathur, and M. L. Soffa, "Automated Test Data Generation using An Iterative Relaxation Method" ACM SIGSOFT Sixth International Symposium on Foundations of Software Engine ering(FSE-6), pages 231-244, Orlando, Florida, November 1998.

No context found.

N. Gupta, A. P. Mathur, and M. L. Soffa. Automated test data generation using an iterative relaxation method. In SIGSOFT '98/FSE-6: Proceedings of the 6th ACM SIGSOFT international symposium on Foundations of software engineering, pages 231-- 244. ACM Press, 1998.

No context found.

N. Gupta, A. P. Mathur, and M. L. Soffa. Automated test data generation using an iterative relaxation method. In SIGSOFT '98/FSE-6: Proceedings of the 6th ACM SIGSOFT international symposium on Foundations of software engineering, pages 231--244. ACM Press, 1998.

No context found.

Neelam Gupta, Aditya P. Mathur, and Mary Lou Soa. Automated Test Data Generation Using an Iterative Relaxation Method. In Foundations of Software Engineering, pages 231-244, 1998.

No context found.

N. Gupta, A. P. Mathur, and M. L. So#a. Automated test data generation using an iterative relaxation method. In ACM SIGSOFT Sixth International Symposium on Foundations of Software Engineering(FSE-6), Nov. 1998.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC