Coverage and Usage Testing
Based on FSMs and Markov Chains
FSM-Based Testing
FSM:
- State: operations/ functions
- Transition: link int a chain
- Input/output associated with transition
- Complete operation: chain
Three representations of FSMs
- Graphical
- Tabular
- list
- define state
{Power-Up, Mobile Station Initialization, Mobile Station Idle, System access, Mobile station control or Traffic channel}
{A, B, C, D, E}
- define transitions
{A -> B; B-> C; C->D (paging channel message); C-> B (unable to receive paging channel); D->C (finished other tasks); D -> E (making a call); E -> B}
Graphical
Tabular
- define row and col meaning
The rows represent originating states and the columns represent ending states for specific transitions.
- define cell
A null input or a null output is marked by “-”.
msg , NoC , call, and done to represent “paging channel message”, “unable to receive paging channel”, “making a call”, and “finished with other tasks”, respectively.
“na” or not marked (left empty), the corresponding transition is not allowed.
List
{A, B, - , -}
{B, C, -, -}
{C, B, “unable to receive paging channel”, -}
{C, D, "paging channel message", -}
{D, C, "finished with other tasks", -}
{D, E, "making a call", -}
{E, B, -, -}
Markov Chains as Enhanced FSMs
FSM+发生的概率 = Markov Chains
In the context that we use it for usage-based statistical testing, this Markov chain is also called a Markov OP, because it constitutes the specific operational profile (OP) for the system.
Musa’s operational profiles (OPs) for usage-based statistical testing using partitions.
Starting with a simple Markov OP example of your choice, convert it to the corresponding Musa OP. (Make necessary assumptions as needed.)
output sum should be 1
Musa OP: an operational profile, or an OP for short, is a list of disjoint set of operations and their associated probabilities of occurrence.
Unified Markov Models for Testing
统一马尔可夫模型
这样的 UMMs 使得模拟个别操作单元并将它们链接起来形成全局操作变得简单。更高层次的操作可以扩展到更低层次的模型中,以进行更彻底的测试。