| Summary: | Submitted by Dilene Paulo (dilene.fatima@ufcg.edu.br) on 2018-01-25T12:53:52Z
No. of bitstreams: 1
ADRIANA CARLA DAMASCENO – TESE PPGCC 2015.pdf: 3272500 bytes, checksum: 779024871de692299535f2de0eeabfb9 (MD5) === Made available in DSpace on 2018-01-25T12:53:52Z (GMT). No. of bitstreams: 1
ADRIANA CARLA DAMASCENO – TESE PPGCC 2015.pdf: 3272500 bytes, checksum: 779024871de692299535f2de0eeabfb9 (MD5)
Previous issue date: 2015-03-06 === We can find many examples of Real-time Systems (RTS) in critical applications such
as patient monitoring, air traffic control and others. A failure in this kind of system
can be catastrophic. For example, it can harm human lives or increase project budgets.
Hence, the testing of real-time systems must be accurate. Models are used to
perform this task, since they contain information about how the system behaves and
when actions may happen. Due to the complexity of the available systems, most RTS
are composed of subsystems that interact as part of a bigger system. These subsystems
are combined through operators to model their specification behavior. However, works
on the testing of compositional models for RTS are practically nonexistent. Among the
available approaches to perform testing for non-compositional RTS models, the tioco
conformance testing theory focuses on generating test cases based on implementation
and specification models. Moreover, a conformance relation defines whether success in
testing means conformance between an implementation and a specification. To express
specifications and to represent implementations under test, we use Timed Input Output
Symbolic Transitions Systems (TIOSTS). These models store symbolic data and
clock variables, avoiding the state space and region explosion problems. Regarding the
testing of compositional models, some questions may arise: If two subsystem implementations
are tioco conformant to their specifications, is it correct to assume that the
composition of the implementations is also tioco conformant to the composition of their
specifications? In this case, how can operators be defined to work with TIOSTS and
tioco? To answer these questions, this thesis proposes the sequential, interruption and
parallel operators for the TIOSTS model. For each operator, we study how the tioco
conformance relation behaves with respect to subsystems and the composed system.
We present results towards properties of compositional operators when the subsystems
are composed, as well as implementing them. Besides, we show three examples
where each operator can be used and illustrate the applicability of our approach in two
exploratory studies. The first models components of a aircraft specification and the
second presents application level interruptions in an Android system. === We can nd many examples of Real-time Systems (RTS) in critical applications such as patient monitoring, air tra c control and others. A failure in this kind of system
can be catastrophic. For example, it can harm human lives or increase project budgets.
Hence, the testing of real-time systems must be accurate. Models are used to
perform this task, since they contain information about how the system behaves and
when actions may happen. Due to the complexity of the available systems, most RTS
are composed of subsystems that interact as part of a bigger system. These subsystems
are combined through operators to model their speci cation behavior. However, works
on the testing of compositional models for RTS are practically nonexistent. Among the
available approaches to perform testing for non-compositional RTS models, the tioco
conformance testing theory focuses on generating test cases based on implementation
and speci cation models. Moreover, a conformance relation de nes whether success in
testing means conformance between an implementation and a speci cation. To express
speci cations and to represent implementations under test, we use Timed Input Output
Symbolic Transitions Systems (TIOSTS). These models store symbolic data and
clock variables, avoiding the state space and region explosion problems. Regarding the
testing of compositional models, some questions may arise: If two subsystem implementations
are tioco conformant to their speci cations, is it correct to assume that the
composition of the implementations is also tioco conformant to the composition of their
speci cations? In this case, how can operators be de ned to work with TIOSTS and
tioco? To answer these questions, this thesis proposes the sequential, interruption and
parallel operators for the TIOSTS model. For each operator, we study how the tioco
conformance relation behaves with respect to subsystems and the composed system.
We present results towards properties of compositional operators when the subsystems
are composed, as well as implementing them. Besides, we show three examples
where each operator can be used and illustrate the applicability of our approach in two
exploratory studies. The rst models components of a aircraft speci cation and the
second presents application level interruptions in an Android system.
|
|---|
