KR101606283B1 - Appratus for automatic generation of expected results for testing component-based software system - Google Patents

Abstract

Disclosed are a method and an apparatus for automatically generating an expected result value for testing a component-based software system including: an input unit which receives test data and prediction result values of each of component units which constitute the component-based software system and are connected to each other; a processor which generates an input and output correlation direction graph of the component-based software system, generates a first test data group for testing the component-based software system based on each test data of each component unit, and generates a second test data group which is a subgroup of the first test data group by using the input and output correlation direction graph; and an output unit which generates and outputs a total prediction result value based on the second test data group by using the input and output correlation direction graph. According to the present invention, time and costs for testing the input and output correlation direction graph can be reduced and an accurate test can be performed.

Description

[0001] APPARATUS FOR AUTOMATIC GENERATION OF EXPECTED RESOURCES FOR TESTING COMPONENT-BASED SOFTWARE SYSTEM [0002]

The present invention relates to a method and apparatus for automatically generating an expected result value for testing a component based software system.

A component-based software system is a system composed of two or more software components, each of which is connected via an input / output interface. To test a component-based software system, expected results are needed to verify correct operation, and manually generating these expected results can be time consuming and costly. Conventional technologies have attempted to reduce time and cost by automatically generating expected results for general systems rather than component based software systems. However, the conventional technologies do not take into account the input / output association of the unit components constituting the component-based software system, and have a problem that the expected result value can not be automatically generated when the component-based software system is upgraded.

The embodiments of the present invention analyze the input / output relationship of the composite component using the internal input / output relationship of the unit component and the input / output association information between the unit components, and search the test result of each unit component based on the input / A method and an apparatus for automatically generating an expected result value of an expected result value of an image.

An apparatus for automatically generating an expected result value for testing a component based software system according to an exemplary embodiment of the present invention includes an input unit for receiving test data and expected result values of each of a plurality of interconnected unit components constituting the component- , Generating an input / output association direction graph of the component-based software system, generating a first set of test data for testing the component-based software system based on test data of each of the plurality of unit components, A processor for generating a second set of test data that is a subset of the first set of test data using a relationship direction graph and a processor for generating a second set of test data based on the second set of test data using the input / It may comprise an output for generating the output.

The input / output association relationship direction graph may be generated based on input / output association information within the plurality of unit components and input / output association information between the plurality of unit components.

The second set of test data may be generated based on a determination of whether at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system.

If, as a result of the determination, the at least one output parameter of the component-based software system varies depending on all of the input parameters of the component-based software system, the processor may cause the second test data set to include all of the first test data set . ≪ / RTI >

And if it is determined that each of the output parameters of the component based software system does not depend on all of the input parameters of the component based software system, Generate sets of associated input parameters, and generate the second set of test data based on the set of associated input parameters.

The output may generate a first expected result value that represents an output for the second set of test data and may generate the overall expected result value from the first expected result value.

The output unit may generate the first expected result value based on the test data and the expected result value of each of the plurality of unit components.

An apparatus for automatically generating an expected result value for testing a component based software system according to an exemplary embodiment of the present invention includes an input unit for receiving test data and expected result values of each of a plurality of interconnected unit components constituting the component- Based on the test data of each of the plurality of unit components, the input / output association relationship graph of the component-based software system based on input / output association information between the plurality of unit components, A processor for generating a first test data set for a first test data set and a second test data set that is a subset of the first test data set using the input / output association direction graph, And the first may include a total estimated result output generating and outputting a value based on the second test data set.

The second set of test data may be generated based on a determination of whether at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system.

An apparatus for automatically generating an expected result value for testing a component based software system according to an exemplary embodiment of the present invention includes an input unit for receiving test data and expected result values of each of a plurality of interconnected unit components constituting the component- Based software system; and generating at least one output parameter of the component-based software system based on the test data of each of the plurality of unit components, A processor for generating a second set of test data that is a subset of the first set of test data based on a determination as to whether or not it depends on all of the input parameters, And an output unit for generating and outputting a result value.

Wherein the processor generates an input / output association direction graph of the component-based software system based on input / output association information within the plurality of unit components and / or input / output association information between the plurality of unit components, The test data set and the total expected result value may be generated using the input / output association direction graph, respectively.

If, as a result of the determination, the at least one output parameter of the component-based software system varies depending on all of the input parameters of the component-based software system, the processor may cause the second test data set to include all of the first test data set . ≪ / RTI >

And if it is determined that each of the output parameters of the component based software system does not depend on all of the input parameters of the component based software system, Generate sets of associated input parameters, and generate the second set of test data based on the set of associated input parameters.

The output may generate a first expected result value that represents an output for the second set of test data and may generate the overall expected result value from the first expected result value.

The output unit may generate the first expected result value based on the test data and the expected result value of each of the plurality of unit components.

A method for automatically generating an expected result value for testing a component-based software system according to an embodiment of the present invention includes receiving test data and an expected result value of each of a plurality of interconnected unit components constituting the component- Generating a first test data set for testing the component based software system based on test data of each of the plurality of unit components, Generating a second test data set that is a subset of the first test data set using an input / output association direction graph, and generating a second test data set based on the second test data set based on the input / And generating and outputting a value.

The input / output association relationship direction graph may be generated based on input / output association information within the plurality of unit components and input / output association information between the plurality of unit components.

The second set of test data may be generated based on a determination of whether at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system.

Wherein if the at least one output parameter of the component-based software system is determined to depend on all of the input parameters of the component-based software system, then generating the second set of test data comprises: To include all of the first set of test data.

Wherein if the result of the determination is that each of the output parameters of the component-based software system does not depend on all of the input parameters of the component-based software system, then generating the second set of test data comprises: Generating sets of associated input parameters corresponding to each of the output parameters of the set of associated input parameters, and generating the second set of test data based on the set of associated input parameters.

Generating and outputting an overall expected result value comprises generating a first expected result value representing an output for the second set of test data and generating the overall expected result value from the first expected result value, .

The generating of the first expected result value may include generating the first expected result value based on the test data and the expected result value of each of the plurality of unit components.

A method for automatically generating an expected result value for testing a component-based software system according to an exemplary embodiment of the present invention includes receiving test data and an expected result value of each of a plurality of interconnected unit components constituting the component- Based on the test data of each of the plurality of unit components, generating an input / output association relationship graph of the component-based software system based on input / output association information between the plurality of unit components, Generating a first set of test data for testing of the input / output association relationship direction graph, generating a second set of test data that is a subset of the first set of test data using the input / output association direction graph, By using a test based on the second set of data may include generating a total estimated results to output.

The second set of test data may be generated based on a determination of whether at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system.

The present invention provides a method and apparatus for automatically generating an expected result value for testing a component based software system comprising a plurality of unit components. The present invention can reduce the time and cost required for testing a component-based software system, and accurate testing can be performed. The present invention can also be applied to a large-scale system that is not a component-based system.

1 illustrates a component-based software system in accordance with one embodiment.
2 illustrates an apparatus for automatically generating an expected result value for testing a component-based software system according to an embodiment.
3 is a flowchart illustrating an operation of an apparatus for automatically generating an expected result value for testing a component-based software system according to an exemplary embodiment.
4 illustrates an algorithm for generating an input / output association direction graph of a component-based software system according to an embodiment.
5 illustrates an input / output association relationship graph generated based on input / output association information and unit input / output association information between unit components of a component-based software system according to an exemplary embodiment.
FIG. 6 illustrates an input / output association relationship graph generated using input / output association information between unit components of a component-based software system according to an exemplary embodiment.
7 is a flowchart illustrating an operation of an apparatus for automatically generating an expected result value for testing a component-based software system according to an exemplary embodiment.
8 illustrates a reduced expected result value generation algorithm according to an embodiment.
9 illustrates a portion of a reduced estimated result value generation algorithm according to one embodiment.
FIG. 10 is a diagram showing an example of an input / output association relationship graph generated by using input / output association information and unit input / output association information between unit components of a component-based software system according to an exemplary embodiment, .
11 illustrates a process of generating a reduced estimated result value using an input / output association direction graph generated using input / output association information between unit components of a component-based software system according to an exemplary embodiment.
12 illustrates a component-based software system in accordance with one embodiment.
13 illustrates a graph of input / output association relationships of a component-based software system according to an embodiment.
14 illustrates a component-based software system in accordance with one embodiment.
15 shows a graph of the direction of input / output association of a component-based software system according to an embodiment.
16 illustrates a component-based software system in accordance with one embodiment.
17 illustrates a graph of input / output association relationships of a component-based software system according to an embodiment.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to or limited by the embodiments.

1 is a block diagram illustrating a component-based software system (CBS system) according to an exemplary embodiment of the present invention. As shown in FIG. 1, a component-based software system 100 according to an embodiment may include a plurality of interconnected unit components C ₁ , C ₂ , and C ₃ . Specifically, C _i May refer to an i-th unit component that constitutes a component-based software system, and C ⁱ _in and C ⁱ _out May refer to a set of all input parameters and output parameters of C _i , respectively. x _{i , j} and y _{i , k} may respectively mean the j th input parameter and the k th output parameter of C _i . In particular, EXT _in (C _i ) and EXT _out (C _i ) may refer to a set of input parameters from outside and output parameters to the outside of a component-based software system of C _i , respectively. For example, in the component based software system 100 shown in FIG. 1, C ¹ _in = {x _1,1 , x _1,2, x _1,3 } and EXT _in (C ₁ ) = {x _1,2 , x _2,2 }. Also, a CC may refer to a set of all the component components of a component-based software system, and CC _in and CC _out May each refer to a set of input parameters from outside all of the component-based software system and a set of output parameters to the outside. For example, in the component-based software system 100 shown in FIG. 1, CC = {C ₁ , C ₂ , C ₃ }, CC _in = {x _1,1 , x _1,2, x _2,1, x _2,2 } and CC _out = {y _{1 , 1} , y _{1 , 2,} y _{3 , 1,} y _{3 , 2} }. When C _i ∈ CC, C ⁱ _rp can refer to the input / output association within the C _i component, and RP _intra can refer to all C ⁱ _rp sets. Also, RP _inter can mean a set of I / O associations between components, and RP can mean a set of both component I / O associations and inter-component I / O associations. 1, the output of the C ₂ y _{2, 1} is an input of the ₁ x C _{1, 3.} CBS can be defined as CBS = (CC, RP).

It is assumed that the test cases of each of the unit components (i.e. input test data and expected result values) are given. ATC _i May refer to a set of test cases of an i-th unit component, and STC may refer to a set of test cases of all unit components in a component-based software system. When the size of C ⁱ _in is n, ID _n can refer to a set of input test data. When the size of C ⁱ _out is m, OD _m can mean a set of output data. The set of test cases for the ith unit component is ATC _i = ≪ ID _n , OD _m ≫.

Tables 1 to 3 illustrate the relationship between C ₁ , C ₂ , and C ₁ for the component-based software system 100 shown in FIG. And C ₃ Each is an exemplary test case. A1 through L2 are exemplary input data and output data of each of the unit components. The input test data may be generated using a black-box testing method. In this embodiment, STC = {ATC ₁ , ATC ₂ , ATC ₃ } to be.

Referring to FIG. 2, an apparatus 200 for automatically generating an expected result value for testing a component-based software system 100 according to an embodiment is provided. 2, the expected result automatic generating apparatus 200 may include an input unit 210, a processor 220, and an output unit 230. As shown in FIG. The predicted result value automatic generation apparatus 200 has an input / output association relation in unit components constituting the component-based software system 100, an input / output association relation between unit components, test data of each unit component, Based on which the expected result values for testing of the component based software system 100 can be automatically generated.

Specifically, the input unit 210 may receive test data and expected result values of each of the unit components. In this embodiment, the test data and the expected result value of each of the unit components to which the input unit 210 is input may be a set of test cases (STC) of each of the unit components shown in Tables 1 to 3.

The processor 220 is an I / O relationship relationship graph (I / O relationship) of the component-based software system 100 based on the input / output relationship RP _intra and the input / output relationship RP _inter between the unit components directed graph (IORG). The processor 220 may generate an entire test data set (e.g., a first test data set) of the component based software system 100 based on the input data of each of the unit components. The processor 220 generates a reduced test data set (e.g., a second test data set) from the entire test data set of the component-based software system 100 using the input / output association relationship graph of the component- can do. The reduced test dataset is a subset of the entire test dataset, from which it is generated to derive the overall expected outcome, which is the output for the input of the entire test dataset.

Output 230 may generate a reduced expected result value (e.g., a first expected result value) indicative of an output for an input of the reduced set of test data. The output unit 230 can generate and output the entire expected result value based on the reduced estimated result value.

Hereinafter, the operation of the expected result value automatic generation device 200 for testing the component-based software system 100 according to an embodiment will be described in detail with reference to FIG.

In step 310, the input unit 210 may receive the test data and the expected result value of each of the unit components. In this embodiment, the test data and the expected result value of each of the unit components to which the input unit 210 is input may be a set of test cases (STC) of each of the unit components shown in Tables 1 to 3.

IO affinity directed graph in step 320, processor 220 to component-based software system 100 based on the input-output relation between the unit input and output relationship of the internal components (RP _intra) and the unit components (RP _inter) ( I / O relationship directed graph (IORG). When V is a set of all input and output parameters and E is a set of I / O associations, the I / O association direction graph can be a (V, E) pair.

4 illustrates an algorithm for generating an input / output association direction graph of a component-based software system 100 according to an embodiment. 4, the inputs of the algorithm include a set of both the component internal input / output associations and the input / output associations between components (RP), a set of all external input parameters (CC _in ) of the component-based software system, (CC _out ) of all external output parameters.

The algorithm first initializes V and E of IORG (lines 1 and 2). Then, for each element x of CC _in , we call function GenerateIOReGraph () (line 5) with x as the argument to add x to V of IORG (line 4) and x as the argument to produce V and E associated with x, , Add the result of this function to the IORG (line 6).

The function GenerateIOReGraph () can be constructed as follows. If a set of internal input and output components affinity (RP _intra) is not the empty set and take the set of intra-output associated with the input inp in the set of all the components inside the input-output relationship _{(RP intra) (line 11,12)} . Alternatively, if the set of component internal I / O associations (RP _intra ) is empty, that is, if the internal I / O associations of the unit components are unknown, take all output sets of the corresponding unit component including input inp (line 13, 14,15). For each intra-output rout, add an edge from rout and inp to rout (lines 17, 18). It then takes a set of inter-inputs associated with rout in the set of input / output associations (RP _intra ) between all components (line 19). If the set of inter-inputs is an empty set, or if rout is an element of CC _out , it does not retrieve the associated inputs and outputs (line 20). Otherwise, for each inter-input rin, add the edges from rin and rout to rin (line 22, 23) and recursively call the function GenerateIOReGraph () with line parameter rin (line 24).

5 illustrates an input / output association relationship graph generated by using input / output association information within a plurality of unit components and input / output association information between a plurality of unit components of the component-based software system 100 according to an exemplary embodiment . Referring to FIG. 5, the input / output association direction graph is composed of seven input nodes, seven output nodes, and edges. The input-output association relationship graph may identify each external input parameter of the component-based software associated with each external output parameter of the component-based software system. For example, the external input parameters x _2,1 and _2,2 x is associated to an external output parameter y _{3, 1.}

FIG. 6 shows an I / O association direction graph generated using only input / output association information between a plurality of unit components of the component-based software system 100 according to an embodiment. Referring to FIG. 6, since the input / output association information in the plurality of unit components is not used, a virtual input / output association relation corresponding to the input / output association relationship in the plurality of unit components is generated. Similar to the direction graph, it consists of seven input nodes, seven output nodes, and edges.

Referring again to FIG. 3, in step 330, the processor 220 generates an entire test data set (e.g., a first test data set) of the component based software system 100 based on the input data of each of the unit components . The generation of the entire test data set can be expressed as:

T _in = D (x ₁ ) x D (x ₂ ) x × D (x _p )

Here, T _in denotes a test data set, x _i denotes an i-th input parameter of CC _in , D (x _i ) denotes a set of data that can be input in the input parameter x _i , and p denotes CC _in Quot; size " D (x _i ) is generated in advance using black-box testing and white-box testing techniques in testing unit components. For example, in the component-based software system 100 shown in FIG. 1, D (x1, ₁ ) = {A1, A2}, D (x _1,2 ) = {B1 _{, 1} ) = {F1, F2}, D (x2, ₂ ) = {G1, G2}. Table 4 shows the T _in .

In step 340, the processor 220 uses the I / O association direction graph of the component-based software system 100 to generate a reduced set of test data (e.g., second test data) from the entire set of test data in the component- Set). Here, the reduced set of test data is a subset of the entire set of test data, from which it is generated so as to derive the overall expected result value that represents the output for the input of the entire test data set.

The reduced set of test data may be generated based on the set of associated input parameters corresponding to each of the output parameters. In this specification, T ^red _in may mean a reduced set of test data, and R (y _k ) may refer to a set of associated input parameters corresponding to the output parameter y _k . When the set of input parameters associated with the output parameter y _k is X (y _k ), the set of associated input parameters R (y _k ) can be expressed as:

The expected value of the output parameter y _k is determined by the element of R (y _k ). R (y _k ) can be changed according to the input / output relationship within the component and the input / output association between the components.

X (y _{1 , 1} ) = {x _1,1 , x _1,2 }, X (y _{1 , 2} ) = {x _1,2 , _{x 2,1}, x (y 3} , 1) = a {x _2,1, x _2,2}, and _{x (y 3, 2) =} {x 2,2} are determined. Tables 5 to 8 present combinations of input test data for R (y _{1 , 1} ), R (y _{1 , 2} ), R (y _{3 , 1} ) and R (y _{3 , 2} ), respectively. The symbol "-" means irrelevant data.

A subset of T _in, T _in ^red to cover a union of each R _(k y) can be expressed as follows.

Equation (6)

Referring to Table 9, T ^red _in which the T _{in shown in} Table 4 can be covered .

At step 350, the output 230 may generate a reduced expected result value (e.g., a first expected result value) indicative of an output for an input of the reduced set of test data. In the present specification, T ^red _out may mean a reduced expected result value, and T _out may mean a total expected result value.

8 illustrates a reduced expected result value generation algorithm according to an embodiment. 8, the input of the algorithm may be a reduced test data set (T ^red _in ), an I / O association direction graph (IORG), and a set of test cases (STC) of all the unit components in the component- have.

The algorithm first initializes T ^red _out (line 1). Next, we take the nodes of the CC _{in in} the IORG and the nodes of the CC _out (lines 2 and 3). CI consists of CI.node which stores nodes of CC _in and CI.data which stores elements of T ^red _in . CO consists of CO.node which stores the nodes of the CC _out and CO.data which stores the expected result value of the elements of T ^red _in . For each element inTD of T ^red _in , assign inTD to CI.data and NULL to CO.data (lines 5 and 6). Next, call the function SearchExpectedResults () with the CI and CO as arguments to generate the expected result value for inTD. The expected result value for the generated inTD is stored in CO.data to be added to T ^red _out (line 9).

Figure 9 illustrates the specification of a function SearchExpectedResults () according to one embodiment. 9, the input of this function may be a set of tracing nodes in the IORG and a set of nodes in the CC _out .

The algorithm first groups the tracking nodes according to their associated child nodes (line 3). That is, the tracking nodes in the same group have the same child node. An RNG is a set of groups of tracking nodes with identical child nodes. Next, initialize the next set of tracking nodes nextN (line 4). For each element rn of the RNG, child nodes of rn are assigned to cN.node (line 6), and test cases of unit components associated with rn and cN.node are taken from the STC (line 7). If all nodes of rn are input nodes, take the test cases associated with rn.node and cN.node. Otherwise, take the test cases associated with cN.node only. RTC refers to association test cases of unit components. If rn.data is present in the RTC, the expected result generation fails (line 16). Otherwise, the RTC takes the data associated with cN.node (line 9). If cN.node is in CO.node, cN.data is the expected result to be added to CO.data (line 11). Then, if there is a child node of cN, add cN to nextN. Finally, recursively invoke the function SearchExpectedResults () until nextN is empty (line 19). The time complexity of this algorithm is as follows.

Here, i denotes the depth of the IORG, m denotes the maximum depth of the IORG, RNG (i) denotes a set of groups of tracking nodes having the same child node, RTC (rn) denotes rn May refer to a set of test cases of unit components associated with the unit.

FIG. 10 shows a process of generating a T ^red _out of the component-based software system 100 shown in FIG. Referring to FIG. 10, rn (cN) denotes a group of tracking nodes having the same child node cN. In this embodiment, STC shown in Tables 1 to 3, IORG shown in FIG. 4, and T ^red _in of . _{CI.node = {x 1,1, x 1,2} , x 2,1, x 2,2} _{a, CO.node = {y 1, 1} , y 1, 2, y 3, 1, y 3, ₂ }. Each element of T ^red _in is assigned to CI.data. If r _{1 in} Table 9 is assigned to CI.data, then CI.node is selected as the first tracking nodes. Then, the tracking nodes are assigned to the child nodes y _{1 , 1} , y _{2 , 1,} y _{2 , 2} , y _{2 , and 3} , respectively. Child node y _1, x _1,3 of the parent node ₂ because the tracking node, child node y _1, no trace group are associated with _two nodes. Thus, RNG is _{{rn (y 1, 1)} , rn (y 2, 1), rn (y 2, 2), rn (y 2, 3)}. When rn is rn (y _1,1 ), rn.node is {x _1,1 , x _1,2 } and rn.data is {A1, B1}. Since all nodes of rn (y _1,1 ) are input nodes and rn (y _{1 , 1} ) and cN.node are associated with ATC ₁ , child node y _{1 , 1} of rn (y _{1 , 1} ) node, and then assigns the ATC _{1 shown} in Table ₁ to the RTC. Then, cN.node exists in CO.node and cN.data can be added to CO.data. This is the expected result value for the external output y _{1 , 1} . is also treated in the similar manner with respect to _{rn (y 2,1), rn (} y 2, 2) and rn (y _{2, 3).} Thus, CO.data = {D1}, nextN.node = is a _{{y 2, 1, y 2} , 2, y 2, 3} and nextN.data = {C1, H1, J1 }.

Next, recursively call the function SearchExpectedResults () with the nextN as an argument. RNG is {rn (x _1,3 ), rn (x _3,1 ), rn (x _3,2 )} as shown in Fig. 10 (b). When rn is rn (x _1,3 ), rn.node is {y _{2 , 1} } and rn.data is {C1}. Since all nodes of rn (x _1,3 ) are output nodes and cN.node is associated with ATC ₁ , child nodes x _{1, 3} of rn (x _1,3 ) are assigned to cN.node, allocates ATC ₁ shown in the RTC. As we have seen, if all nodes of rn are output nodes, we assign the test cases of the unit component associated with cN.node to the RTC. Then, since rn.data is present within the ATC ATC ¹ _{1 1} allocates the data C1 to cN.data. Since child node of cN.node is y _{2 , 1} , cN can be added to nextN. Thus, nextN.node = {x _1,3, x _3,1 , x _3,2 } and nextN.data = {C1, H1, J1}.

Finally, we call the function SearchExpectedResults () with nextN as the argument. RNG is the _{{rn (y 1, 2)} , rn (y 3, 1), rn (y 3, 2)} , as shown in 10 (c) Fig. In this step, since the child node y _1, the parent node of the ₂ x _1,1 x _1,3 and is tracking nodes, grouping the nodes associated with the child node trace y _{1, 2.} Since the cN.node of each element of the RNG is in the CO.node, all cN.data are associated with the external outputs y _{1 , 2,} y _{3 ,} y _{3 and 2} , respectively. Therefore, CO.data = {D1, E1, K1, L1} and nextN.node = ?.

As a result of the above procedure, a function SearchExpectedResults () returns the CO.data = {D1, E1, K1 , L1} CI.data when the r ₁ day.

Referring to Table 10, results of applying the algorithms of FIGS. 8 and 9 to the component-based software system 100 shown in FIG. 1 are presented. In Table 10, for each column IORG the depth (depth), associated group of nodes (RNG), the test cases are associated to each group of nodes (RTC), an external output that varies depending on the node groups the node (Rex) and T ^red _out Represents the number of failures in the expected result value generation. Only one expected result value for the external output y _{3 , 1} is not generated. When CI.data is r ₄ in Table 9, the function returns CO.data = {D4, E4, NULL, L2}. Specifically, when rn is rn (x _3,1 ) with depth 3, rn.node is {y _{2 , 2} } and rn.data is {H4}. Assign child node x _3,1 to cN.node, and then assign ATC _{3 shown} in Table ₃ to RTC. Since ATC ₃ does not have rn.data, it deletes all nodes associated with cN.node. That is, when CI.data is r ₄ in Table 9, the generation of the expected result value for y _{3 , 1} fails.

Referring to Table 11, for the ^red T _in of the component-based software system 100 shown in Figure 1 T ^red _out . NULL means that the expected result value was not automatically generated. According to the temporal complexity representation, the time complexity can be calculated as follows.

O (.) = 4 * ((8 + 4 + 4 + 4) + (8 + 6 + 6) + (8 + 6 + 6)

= RNG (3) | = 3, | ATC ₁ | = 8, | ATC ₁ | = 4, | RNG (2) | = 3, | T ^red _in | = 4, m = ₂ | = 4, | ATC ₃ | = 6.

In Table 11. Estimated results for the external output y _3, r ₁ of ₄ K9 can be determined by a value manually.

11 illustrates a process of generating a T ^red _out of the component-based software system 100 shown in FIG. 1 when the input / output association within a plurality of unit components of the component-based software system 100 is unknown. Referring to FIG. 11, an input / output association direction graph generated using only input / output association information between a plurality of unit components of the component-based software system 100 may be used. Since the input / output association information in the plurality of unit components is not used, a virtual input / output association corresponding to the input / output association in the plurality of unit components is generated, and a similar method as described with reference to FIG. 10 To generate a T ^red _out .

Referring back to FIG. 3, in step 350, the output unit 230 may generate and output the entire expected result value based on the reduced estimated result value. When T ^red _out is generated from the T ^red _out T _out Can be expanded. X _D (y _k ) may mean a set of input test data X (y _k ) _in T _in . First, each element _in T _in is separated according to X (y _k ). For example, for the element t ₁ of Table 4, the element _{X D (y 1, 1)} = {A1, B1}, X D (y 1, 2) = {B1, F1}, X D (y 3 _, may be isolated, such as _{1) = {F1, G1}} , X D (y 3, 2) = {G1}. Then, the expected result for each X _D (y _k) is retrieved from the T ^red _out to the T _in ^red. Table 12 presents T _out for the T _in of the component-based software system 100 shown in FIG.

Hereinafter, the operation of the expected result value automatic generation device 200 for testing the component-based software system 100 according to an exemplary embodiment will be described in detail with reference to FIG. The operation of the expected result value automatic generation apparatus for the composite component test shown in FIG. 7 can be applied as it is with reference to FIG. 3. Therefore, the difference from the operation shown in FIG. 3 will be mainly described.

At step 740, the processor 220 may determine whether at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system. If, if any one output of the component-based software system changed depending on all of the input parameters (i. E., Are identical and some X (y _k) is CC _in), the since the T ^red _in = T _in T ^red _in You do not need to create a new one. Otherwise, T ^red _in may be a smaller set than T _in . The possible size of T ^red _in can be expressed as:

_{Max (| R (y 1)} |, ..., | R (y q) |) ≤ | T red in | ≤ | T in |

In step 750, if it is determined that at least one output parameter of the component-based software system dalrajindago in dependence on all of the input parameters of the component-based software system, the processor 220 is the same T _in and T _in ^red Can be generated.

In step 760, if it is determined that each of the output parameters of the component-based software system does not depend on all of the input parameters of the component-based software system, the processor 220 may correspond to each of the output parameters of the component- Lt; RTI ID = 0.0 > input < / RTI > The set of associated input parameters may be generated in a manner similar to that in step 340 of FIG.

At step 770, the processor 220 generates a T ^red _in less than T _in based on the generated set of associated input parameters Can be generated. T ^red _in May be generated in a similar manner as in step 340 of FIG.

12 illustrates a component-based software system in accordance with one embodiment. Referring to FIG. 12, the component-based software system 1200 has a modified input / output relationship with respect to the component-based software system 100. In such a case, as shown in FIG. 12, an expected result value can be generated by searching only the input / output relationship associated with the changed external output. Component-based software system 1200 is an input parameter x of the C ₂ and C ₃ of the output parameter, y _{2,4, 3,3} has been added as compared to component-based software system 100, the external output parameter y _3, the expected result of ₂ The value can be changed. Therefore, the output parameter y _3, the nodes associated with the _second (i.e., x _2,2, y _2,3, y _{2, 4,} x _3,2, and x _3,3) are grouped, only, T ₃ y in the ^red _{out , The} test cases associated with the groups are retrieved again to produce an expected result value of ₂ . Finally, the modified T ^red _out can again be extended to T _out .

Hereinafter, experimental data for verifying effects according to embodiments of the present invention is presented. _{Assuming that the} cost of manually generating the expected value is C _er , C _er Can be expressed as follows.

C _er = E _m / E _t

Where E _m may refer to the number of expected result values generated manually, and E _t may refer to the total number of all expected result values generated. The cost C _er increases as the expected outcome value, which must be determined manually. In order to verify the effect of the embodiments according to the present invention, the measured C _er in each example is referred to as "PJ Schroeder, P. Faherty, B. Korel, Generating expected results for automated black-box testing, in: Automated Software Engineering , 2002.Proceedings. ASE 2002. 17th IEEE International Conference on, IEEE, 139-148, 2002 ( hereinafter, 'Schroeder et al.' referred to) are shown in comparison to C according to the _er ".

I. Example  One

According to this embodiment, the effect of the invention of the external output of the CC _out each case differ depending only on a part of the external input of the CC _in (that is, does not vary in dependence on all of the external inputs of CC _in) Can be verified. In the present embodiment, T ^red _in may be smaller than T _in .

In this embodiment, a component-based software system 1400 as shown in Fig. 14 is used. The component based software system 1400 avoids obstacles, moves the robot, and generates a guidance message. The component based software system 1400 includes a distance sensor component C _dsc , a moving component C _mc , a message generating component C _sgc , and a control component C _cc .

Referring to FIG. 15, there is shown an I / O association direction graph (IORG) of a component based software system 1200. The input / output association direction graph of the component based software system 1400 is generated by the algorithm shown in FIG. According to the IORG in FIG. 15, each external output of the component based software system 1200 depends on only some of the external inputs as follows.

X (sgc_out1) = {dsc_in1, dsc_in2, dsc_in3, dsc_in7, sgc_in1}

X (mc_out1) = {dsc_in4, dsc_in5, dsc_in6, dsc_in8}

For the performance of the experiment, the test cases of each of the unit components (i. E. Input test data and expected result values) are given. Tables 13 through 16 are input parameter specifications and test cases of the distance sensor component C _dsc , the control component C _cc , the message generation component C _sgc , and the mobile component C _mc , respectively.

For each unit component, input test data of various strengths are combined. The minimum and maximum strengths are determined by the number of input parameters of each unit component. Then, an expected result value of the combination of the input test data of each unit component is generated. Table 17 shows the set of test cases for all of the unit components. t means the strength at the time of generation of the ATC, and the symbol "- " means that the number of input parameters is unity because the number of input parameters is one. For example, when ATC _mc is generated, input test data is not combined because C _mc has only one input parameter (mc_in1). For example, STC ₇ consists of ATC _dsc generated at t = 8, ATC _cc generated at t = 3, ATC _sgc generated at t = 3, and ATC _mc .

Next, a T _in of the component-based software system 1200 is generated. Table 18 provides a specification of the external input parameters of the component-based software system 1200 and its input test data.

To generate T _in , the input test data is combined as follows.

_{T in = D (dsc_in1) ×} D (dsc_in2) × D (dsc_in3) × D (dsc_in4) × D (dsc_in5) × D (dsc_in6) × D (dsc_in7) × D (dsc_in8) × D (sgc_in1),

| T _in | = 4 * 4 * 4 * 3 * 3 * 3 * 2 * 2 * 2 = 13824.

Next, T ^red _in which the upper T _in can be covered . For each external output (sgc_out, mc_out), a set of combinations of input test data associated with each external output is generated.

R (sgc_out1) = D (dsc_in1) D (dsc_in2) D (dsc_in3) D (dsc_in7) D (sgc_in1)

R (mc_out1) = D (dsc_in4) D (dsc_in5) D (dsc_in6) D (dsc_in8)

| R (sgc_out1) | = 4 * 4 * 4 * 2 = 256,

| R (mc_out1) | = 3 * 3 * 3 * 2 = 54.

Therefore, T ^red _in Is generated as follows.

T ^red _in = R (sc_out1) ∪ R (mc_out1),

| T ^red _in | = max (R (sc_out1), R (mc_out1)) = 256.

The total number of expected outcome values of T _out and T ^red _out can be calculated as follows.

| T _out | = | T _in | * | CC _out | = 13824 * 2 = 27648,

| T ^red _out | = | T ^red _in | * | CC _out | = 256 * 2 = 512.

Then, according to the algorithm shown in Figs. 7 and 8 generate a T ^red _out, and expand in a manner similar to that described in Schroeder et al. T ^red _out to T _out.

Table 19 shows the cost C _er of the manual generation of the expected result value at E _t = T _out . The cost of Schroeder et al. Is 0.018, regardless of the set of test cases for all of the component components. In other words, it will automatically expand to T _out after all of the 512 expected results out of T ^red _out manually decide to.

The cost according to this embodiment is in the range of 0.001 to 0.015 depending on the set of test cases for all of the unit components. Using a set of test cases in the 95 to 480 of the 512 estimated result of the T ^red _out because the auto-generated, and the remaining 32 to 417 expected results only manually determined.

II . Example  2

According to the present embodiment, the effect of the present invention in the case where the external output of CC _out varies depending on all the external inputs of CC _in can be verified. In this embodiment, T ^red _in may be equal to T _in .

In this embodiment, a component based software system 1600 as shown in FIG. 16 is used. The component-based software system 1600 performs navigation from the origin to the destination, avoids obstacles, moves the robot, and generates a guidance message. The component based software system 1600 includes a global map component C _gmc , a path planning component C _ppc , an obstacle avoidance component C _oac , a navigation component C _nc , a localization component C _lc , (C _bas), and a user detection component (C _udc), the laser scanner component (C _lsc), the mobile component (C _mc) and a message generating component (C _sgc).

Referring to FIG. 17, an input / output affinity direction graph (IORG) of a component based software system 1600 is shown. The input / output association direction graph of the component based software system 1600 is generated by the algorithm shown in FIG. According to the IORG in Figure 17, each external output of the component based software system 1600 depends on all of the external inputs as follows.

X (sgc_out1) = CC _in ,

X (mc_out1) = {gmc_in1, lsc_in1, nc_in1, nc_in2, bsc_in1, udc_in1, udc_in2, udc_in3, udc_in4}.

For the performance of the experiment, the test cases of each of the unit components (i. E. Input test data and expected result values) are given. Tables 20 to 29 illustrate the global map component C _gmc , the laser scanner component C _lsc , the bumper sensor component C _bass , the user detection component C _dc , the localization component C _lc , the navigation component C _nc ), an obstacle avoidance component (C _oac ), a path planning component (C _ppc ), a message generating component (C _sgc ) and a moving component (C _mc ).

As in the previously described embodiment, the input test data is combined for each unit component. Then, an expected result value of the combination of the input test data of each unit component is generated. Table 30 shows the set of test cases for all of the unit components.

Next, a T _in of the component-based software system 1400 is generated. Table 31 provides a specification of the external input parameters of the component-based software system 1400 and its input test data.

To generate T _in , the input test data is combined as follows.

_{T in = D (gmc_in1) ×} D (lsc_in1) × D (nc_in1) × D (nc_in2) × D (bsc_in1) × D (udc_in1) × D (udc_in2) × D (udc_in3) × D (udc_in4) × D ( sgc_in1),

| T _in | = 1 * 6 * 2 * 2 * 2 * 4 * 4 * 4 * 2 * 2 = 12288.

Next, T ^red _in which the upper T _in can be covered . For each external output (sgc_out, mc_out), a set of combinations of input test data associated with each external output is generated.

D (udc_in1) D (udc_in2) D (udc_in3) D (udc_in4) D (gcc_in1) D (lsc_in1) D (nc_in1) D (nc_in2) D (bsc_in1) D (sgc_in1),

(Mc_out1) = D (gmc_in1) 占 D (lsc_in1) 占 D (nc_in1) 占 D (nc_in2) 占 D (bsc_in1) 占 D (udc_in1) 占 D (udc_in2) 占 D (udc_in3) 占 D

| R (sgc_out1) | = | T _in | = 12288,

| R (mc_out1) | = 1 * 6 * 2 * 2 * 2 * 4 * 4 * 4 * 2 = 6144.

Here, since the _{X (sgc_out1) = CC in R} (sgc_out1) = T in. Therefore, T ^red _in is not newly generated, and T ^red _in = T _in . The total number of expected outcome values of T _out and T ^red _out can be calculated as follows.

| T _out | = | T ^red _out | = | T _in | * | CC _out | = 12288 * 2 = 24576.

Table 32 shows the cost C _er of the manual generation of the expected result value when E _t = T _out . The cost of Schroeder et al. Is 1.0 regardless of the set of test cases for all of the component components. That is, all the expected result values of T _out are determined manually.

The cost according to this embodiment is in the range of 0.590 to 0.998 depending on the set of test cases of all the unit components. Using a set of test cases for 35 to 10 052 pieces of the predicted result value T 24,576 pieces of ^red _out because the auto-generated, the remaining 14 524 to 24 541 with only manual determination of expected results.

As shown in the above experimental data, although there is a difference according to the input / output relationship and the test case of the component-based software system, according to the embodiment of the present invention, more expected results are automatically Can be generated. This difference, Schroeder et al. In due to its ability to automatically create the present invention, while generating the expected result of the T ^red _out by always manually by searching in the test case.

Although the preferred embodiments of the present invention have been disclosed for illustrative purposes, those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims. For example, it should be understood that the techniques described may be performed in a different order than the described methods, or that components of the described systems, structures, devices, circuits, and the like may be combined or combined in other ways than the described methods, Lt; / RTI > can be achieved, even if it is replaced or replaced by < RTI ID = 0.0 >

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (23)

An apparatus for automatically generating an expected result value for testing a component-based software system,
An input unit for receiving test data and expected result values of each of a plurality of interconnected unit components constituting the component-based software system;
Output association relationship corresponding to an input / output association relationship in the plurality of unit components, generates an input / output association direction graph of the component-based software system based on the virtual input / output association, Based software system based on the test data of each of the first test data sets and the second test data set, which is a subset of the first test data set, using the input / output association direction graph, Lt; / RTI > And
An output unit for generating and outputting a total expected result value based on the second test data set using the input / output relationship direction graph;
Lt; / RTI >
Wherein the second set of test data is generated based on a determination as to whether at least one output parameter of the component based software system depends on all of the input parameters of the component based software system Automated generation of expected results. The method according to claim 1,
Wherein the input / output association relationship direction graph is generated based on the virtual input / output association information and input / output association information between the plurality of unit components.
An apparatus for automatically generating an expected result value for testing a component based software system. delete The method according to claim 1,
If, as a result of the determination, at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system,
Wherein the processor is configured to cause the second set of test data to include all of the first set of test data.
An apparatus for automatically generating an expected result value for testing a component based software system. The method according to claim 1,
If it is determined that each of the output parameters of the component-based software system does not depend on all of the input parameters of the component-based software system,
Wherein the processor is configured to generate associative input parameter sets corresponding to each of the output parameters of the component based software system and to generate the second set of test data based on the associative input parameter sets.
An apparatus for automatically generating an expected result value for testing a component based software system. The method according to claim 1,
Wherein the output generates a first expected result value that represents an output for the second set of test data and generates the overall expected result value from the first expected result value,
An apparatus for automatically generating an expected result value for testing a component based software system. The method according to claim 6,
Wherein the output unit generates the first expected result value based on the test data and the expected result value of each of the plurality of unit components,
An apparatus for automatically generating an expected result value for testing a component based software system. delete delete An apparatus for automatically generating an expected result value for testing a component-based software system,
An input unit for receiving test data and expected result values of each of a plurality of interconnected unit components constituting the component-based software system;
Generating a first test data set for testing the component-based software system based on test data of each of the plurality of unit components, and generating a virtual input / output association relationship corresponding to the input / output association within the plurality of unit components Based software system; determining whether at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system; and determining, based on the virtual input / output association, A processor for generating a second set of test data; And
An output unit for generating and outputting a total expected result value based on the second set of test data;
And an expected result value automatic generation device for testing a component based software system. 11. The method of claim 10,
Wherein the processor generates an input / output association relationship graph of the component-based software system based on input / output association information between the plurality of unit components and the virtual input / output association, Are generated using the input / output association direction graph, respectively.
An apparatus for automatically generating an expected result value for testing a component based software system. delete 11. The method of claim 10,
Wherein the output generates a first expected result value that represents an output for the second set of test data and generates the overall expected result value from the first expected result value,
An apparatus for automatically generating an expected result value for testing a component based software system. 14. The method of claim 13,
Wherein the output unit generates the first expected result value based on the test data and the expected result value of each of the plurality of unit components,
An apparatus for automatically generating an expected result value for testing a component based software system. A method for automatically generating an expected result value for testing a component based software system,
Receiving test data and expected result values of each of a plurality of interconnected unit components constituting the component-based software system;
Generating a virtual input / output association relationship corresponding to an input / output association relationship in the plurality of unit components, and generating an input / output association direction graph of the component based software system based on the virtual input / output association;
Generating a first set of test data for testing of the component based software system based on test data of each of the plurality of unit components;
Generating a second test data set that is a subset of the first test data set using the input / output association direction graph; And
Generating and outputting an overall expected result value based on the second test data set using the input / output relationship direction graph
Lt; / RTI >
Wherein the second set of test data is generated based on a determination as to whether at least one output parameter of the component based software system depends on all of the input parameters of the component based software system, A method for automatically generating an expected result value. 16. The method of claim 15,
Wherein the input / output association relationship direction graph is generated based on the virtual input / output association information and input / output association information between the plurality of unit components.
Automated generation of expected result values for testing of component - based software systems. delete 16. The method of claim 15,
If, as a result of the determination, at least one output parameter of the component-based software system depends on all of the input parameters of the component-based software system,
Wherein generating the second set of test data comprises allowing the second set of test data to include all of the first set of test data.
Automated generation of expected result values for testing of component - based software systems. 16. The method of claim 15,
If it is determined that each of the output parameters of the component-based software system does not depend on all of the input parameters of the component-based software system,
Wherein generating the second set of test data comprises generating associative input parameter sets corresponding to each of the output parameters of the component based software system and generating the second set of test data based on the associative input parameter sets. ≪ / RTI >
Automated generation of expected result values for testing of component - based software systems. 16. The method of claim 15,
Generating and outputting an overall expected result value comprises generating a first expected result value representing an output for the second set of test data and generating the overall expected result value from the first expected result value, Including,
Automated generation of expected result values for testing of component - based software systems. 21. The method of claim 20,
Wherein generating the first expected result value comprises generating the first expected result value based on test data and an expected result value of each of the plurality of unit components.
Automated generation of expected result values for testing of component - based software systems. A method for automatically generating an expected result value for testing a component based software system,
Receiving test data and expected result values of each of a plurality of interconnected unit components constituting the component-based software system;
Output associating relationship corresponding to an input / output associating relationship in the plurality of unit components, and generating an input / output associating relationship between the input / output associating relationship of the component-based software system and the input / Generating a relationship direction graph;
Generating a first set of test data for testing of the component based software system based on test data of each of the plurality of unit components;
Generating a second test data set that is a subset of the first test data set using the input / output association direction graph; And
Generating and outputting an overall expected result value based on the second test data set using the input / output relationship direction graph
Lt; / RTI >
Wherein the second set of test data is generated based on a determination as to whether at least one output parameter of the component based software system depends on all of the input parameters of the component based software system, A method for automatically generating an expected result value.

delete

Images